Cascaded CT data target recognition method and apparatus, and ray scanning detection system

Abstract

Provided is a cascaded CT data target recognition method, the method comprising: acquiring three-dimensional CT data; performing initial identification on the three-dimensional CT data to obtain a first object description set of a target; and according to the first object description set of the target, using a cascaded fine recognition method to perform fine recognition on the target, so as to obtain a fine recognition result of the target, wherein according to the cascaded fine recognition method, fine recognition is performed on the target by using at least one of a first deep learning network for three-dimensional CT data and a second deep learning network for two-dimensional data, so as to obtain the fine recognition result of the target.

Description

Cascade CT data target recognition method and device, and ray scanning detection system

[0001] Cross-reference to Related Applications

[0002] This application claims the benefit of Chinese Patent Application No. 202411230854.9, filed September 3, 2024, and also claims the benefit of Chinese Patent Application No. 202411231083.5, filed September 3, 2024, and further claims the benefit of Chinese Patent Application No. 202411230846.4, filed September 3, 2024, the disclosures of which are incorporated by reference herein in their entireties. TECHNICAL FIELD

[0003] The present disclosure relates to the technical field of image processing and / or target detection, and in particular to a cascade CT data target recognition method and device, an electronic device, a ray scanning detection system, a computer readable storage medium, and a program product. BACKGROUND

[0004] In related technologies, security CT (Computed Tomography) devices can provide three-dimensional images with more abundant information and are widely used in luggage security checks in customs, airports, train stations, subway stations, and other places.

[0005] With the development of technology, the accuracy and real-time requirements of target recognition in the security field are becoming higher and higher. On the one hand, three-dimensional CT data contains more abundant information, and on the other hand, three-dimensional CT data usually contains a large number of voxels, which are sparse, and also contain certain noise and redundant information. Identifying three-dimensional CT targets will face the problem of huge computational load. In the security scene, it is more difficult to identify targets with strong three-dimensional shape features but complex material composition and physical properties, and the real-time and accuracy requirements of three-dimensional CT data target recognition in the security scene are also higher.

[0006] Therefore, how to effectively use three-dimensional CT data to identify targets is one of the key issues to improve the accuracy and real-time of the algorithm.

[0007] It should be noted that the above information disclosed in this section is only for understanding the background of the inventive concept of the present disclosure, and therefore, the above information can contain information that does not constitute prior art. SUMMARY

[0008] In view of at least one aspect of the above technical problems, a cascaded CT data target recognition method and device, electronic equipment, ray scanning detection system, computer readable storage medium and program product are provided.

[0009] In one aspect, a cascaded CT data target recognition method is provided, the method comprising: acquiring three-dimensional CT data; performing preliminary identification on the three-dimensional CT data to obtain a first object description set of a target; and performing fine identification on the target according to the first object description set of the target by using a cascaded fine identification method to obtain a fine identification result of the target, wherein in the cascaded fine identification method, at least one of a first deep learning network for three-dimensional CT data and a second deep learning network for two-dimensional data is selectively used to perform fine identification on the target to obtain the fine identification result of the target.

[0010] According to some exemplary embodiments, the first object description set comprises one or more of three-dimensional voxel data, point cloud data, feature vectors and semantic descriptions.

[0011] According to some exemplary embodiments, the semantic description comprises at least one of class information, position information and confidence of the target.

[0012] According to some exemplary embodiments, the preliminary identification on the three-dimensional CT data to obtain the first object description set of the target specifically comprises: processing the three-dimensional CT data using a three-dimensional segmentation network based on deep learning to obtain the first object description set of the target; or processing the three-dimensional CT data using a three-dimensional target detection network to obtain the first object description set of the target; or processing the three-dimensional CT data in a target feature library comparison manner to obtain the first object description set of the target.

[0013] According to some exemplary embodiments, the preliminary identification on the three-dimensional CT data to obtain the first object description set of the target specifically comprises: a slice region extraction step of acquiring data of one tomographic section of an object to be detected; pre-processing the tomographic section data; segmenting the pre-processed tomographic section data into a plurality of regions with similar physical properties; in response to a certain segmented region being a slice region, extracting the slice region; a slice object formation step of judging whether the extracted slice region in the current tomographic section can be merged with the extracted slice region in the previous tomographic section; in response to the extracted slice region in the current tomographic section being able to be merged with the extracted slice region in the previous tomographic section, merging to form a slice object; and repeatedly performing the slice region extraction step and the slice object formation step to process the data of each tomographic section by using a tomographic section-by-tomographic section analysis method until all the tomographic section data is processed, to obtain the first object description set of the target.

[0014] Optionally, the preliminary identification of the three-dimensional CT data to obtain the first object description set of the target comprises: a block region extraction step of obtaining data of one slice of the detected object; pre-processing the slice data; segmenting the pre-processed slice data into a plurality of regions with similar physical properties; in response to a segmented region being a block region, extracting the block region; a block object formation step of judging whether the extracted block region in the current slice can be merged with the extracted block region in the previous slice; in response to the extracted block region in the current slice being able to be merged with the extracted block region in the previous slice, merging to form a block object; and repeatedly performing the block region extraction step and the block object formation step to process the data of each slice by using a slice-by-slice analysis method until all the slice data is processed to obtain the object description set of the block target.

[0015] According to some exemplary embodiments, the fine identification of the target according to the first object description set of the target by using the cascaded fine identification method to obtain the fine identification result of the target comprises: inputting the first object description set into the first deep learning network, and processing the first object description set by using the first deep learning network to obtain the fine identification result of the target.

[0016] According to some exemplary embodiments, the method further comprises: generating a second object description set according to the first object description set of the target, wherein the first object description set is an object description set for three-dimensional data, and the second object description set is an object description set for two-dimensional data.

[0017] According to some exemplary embodiments, the fine identification of the target according to the first object description set of the target by using the cascaded fine identification method to obtain the fine identification result of the target comprises: inputting the second object description set into the second deep learning network, and processing the second object description set by using the second deep learning network to obtain the fine identification result of the target.

[0018] According to some exemplary embodiments, in the cascaded fine identification method, the first deep learning network and the second deep learning network are used to perform fine identification on the target to obtain the fine identification result of the target.

[0019] According to some exemplary embodiments, the fine identification of the target by using the first deep learning network and the second deep learning network comprises: inputting the first object description set into the first deep learning network and processing the first object description set by using the first deep learning network; and inputting the second object description set into the second deep learning network and processing the second object description set by using the second deep learning network.

[0020] According to some exemplary embodiments, the inputting the first object description set into the first deep learning network comprises: inputting one or more of three-dimensional voxel data, point cloud data, feature vectors and semantic descriptions corresponding to the target in the first object description set into the first deep learning network.

[0021] According to some exemplary embodiments, the input of the first deep learning network further comprises: background region information in three-dimensional CT image data; and / or, the three-dimensional CT data comprises first three-dimensional image obtained by a first CT detection device; and the input of the first deep learning network further comprises second three-dimensional data obtained by the first CT detection device or other detection devices different from the first CT detection device, wherein the second three-dimensional data comprises second three-dimensional image, the second three-dimensional image corresponds to the first three-dimensional image, and information in the first object description set of the target is mapped into the second three-dimensional image.

[0022] According to some exemplary embodiments, the second object description set comprises one or more of slice data of a three-dimensional region of interest of the target along any direction, feature vectors and semantic descriptions corresponding to the target in the slice data; and / or, the second object description set comprises one or more of two-dimensional projection image obtained by projecting the three-dimensional region of interest of the target, feature vectors and semantic descriptions corresponding to the target in the two-dimensional projection image; and / or, the second object description set comprises one or more of two-dimensional projection image obtained by projecting the three-dimensional CT data as a whole, feature vectors and semantic descriptions corresponding to the target in the two-dimensional projection image; and / or, the three-dimensional CT data comprises first three-dimensional image obtained by a first CT detection device; and the second object description set comprises one or more of two-dimensional image obtained by the first CT detection device or other detection devices different from the first CT detection device, feature vectors and semantic descriptions corresponding to the target in the two-dimensional image.

[0023] According to some exemplary embodiments, the processing the first object description set by using the first deep learning network specifically comprises: performing feature extraction by using the pre-trained first deep learning network to obtain a feature map.

[0024] According to some exemplary embodiments, the processing the second object description set by using the second deep learning network specifically comprises: performing feature extraction by using the pre-trained second deep learning network to obtain a feature map.

[0025] According to some exemplary embodiments, the extracted features comprise at least one of shape, texture and color in the three-dimensional CT data.

[0026] According to some exemplary embodiments, the processing the first object description set by using the first deep learning network specifically comprises: performing operations on the feature map by using the first deep learning network, wherein the operations comprise at least one of feature map fusion and feature map dimension reduction.

[0027] According to some exemplary embodiments, the processing the second object description set by using the second deep learning network specifically comprises: performing operations on the feature map by using the second deep learning network, wherein the operations comprise at least one of feature map fusion and feature map dimension reduction.

[0028] According to some exemplary embodiments, the processing the first object description set by using the first deep learning network further specifically comprises: applying the feature map to a classification, segmentation or detection task to obtain a three-dimensional image semantic description set.

[0029] According to some exemplary embodiments, the processing the second object description set by using the second deep learning network further specifically comprises: applying the feature map to a classification, segmentation or detection task to obtain a two-dimensional image semantic description set.

[0030] According to some exemplary embodiments, the fine recognition of the target by using the cascaded fine recognition method to obtain the fine recognition result of the target specifically comprises: performing the step of processing the input by using the first deep learning network for a single time; and taking the output after the single-time execution of the first deep learning network as the fine recognition result of the target.

[0031] According to some exemplary embodiments, the fine recognition of the target by using the cascaded fine recognition method to obtain the fine recognition result of the target specifically comprises: performing the step of processing the input by using the second deep learning network for a single time; and taking the output after the single-time execution of the second deep learning network as the fine recognition result of the target.

[0032] According to some exemplary embodiments, the target is finely identified by using the cascaded fine identification method to obtain the fine identification result of the target, specifically comprising: iteratively performing the step of processing the input by the first deep learning network; and taking the output after the first deep learning network is iteratively executed multiple times as the fine identification result of the target.

[0033] According to some exemplary embodiments, the target is finely identified by using the cascaded fine identification method to obtain the fine identification result of the target, specifically comprising: iteratively performing the step of processing the input by the second deep learning network; and taking the output after the second deep learning network is iteratively executed multiple times as the fine identification result of the target.

[0034] According to some exemplary embodiments, the fine identification result of the target is in units of one or more of voxels, three-dimensional regions of interest and three-dimensional CT images, and contains: class information and / or confidence; or the fine identification result of the target is in units of three-dimensional regions of interest and / or three-dimensional CT images, and contains: at least one of class information, confidence and position information of the target.

[0035] In another aspect, a cascaded CT data target identification device is provided, comprising: an acquisition module configured to acquire three-dimensional CT data; a preliminary identification module configured to preliminarily identify the three-dimensional CT data to obtain a first object description set of a target; and a fine identification module configured to finely identify the target by using a cascaded fine identification method according to the first object description set of the target to obtain a fine identification result of the target, wherein in the cascaded fine identification method, at least one of a first deep learning network for three-dimensional CT data and a second deep learning network for two-dimensional data is used to finely identify the target to obtain the fine identification result of the target.

[0036] In yet another aspect, an electronic device is provided, comprising: one or more processors; a memory storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method as described above.

[0037] In yet another aspect, a radiation scanning detection system is provided, comprising: a radiation source configured to emit radiation; a detector configured to receive radiation emitted from the radiation source and passing through a detection object; and a data processing device in communication connection with the detector, wherein the data processing device comprises: one or more processors; a memory storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method as described above.

[0038] In still another aspect, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method as described above.

[0039] In still another aspect, there is provided a computer program product comprising a computer program which, when executed by a processor, implements the steps of the method as described above. BRIEF DESCRIPTION OF DRAWINGS

[0040] For a better understanding of the present application, it will now be described in detail in accordance with the following drawings:

[0041] FIG. 1 is a schematic diagram of a structure of a ray scanning detection system according to some example embodiments of the present application.

[0042] FIG. 2 is a schematic diagram of a CT scanning apparatus implementing a reconstruction scanning mode according to some example embodiments of the present application.

[0043] FIG. 3 shows an example of attenuation signal data obtained according to an embodiment of the present application.

[0044] FIG. 4A is a schematic flowchart of a cascaded CT data target recognition method according to some example embodiments of the present application.

[0045] FIG. 4B is a schematic flowchart of a cascaded CT data target recognition method according to some other example embodiments of the present application.

[0046] FIG. 5 is a schematic flowchart of a preliminary recognition operation in a cascaded CT data target recognition method according to an embodiment of the present application.

[0047] FIG. 6 is a schematic flowchart of a refined recognition operation in a cascaded CT data target recognition method according to an embodiment of the present application.

[0048] FIG. 7 is an architectural diagram of a cascaded deep learning network used in a target recognition method according to an example embodiment of the present application.

[0049] FIG. 8 is a schematic flowchart of a cascaded CT data target recognition method according to some other embodiments of the present application.

[0050] FIG. 9 is a structural block diagram of a cascaded CT data target recognition apparatus according to an example embodiment of the present application.

[0051] FIG. 10 schematically shows a structural block diagram of an electronic device suitable for implementing the target recognition method according to an example embodiment of the present application. DETAILED DESCRIPTION

[0052] Specific embodiments of the present application will now be described in detail with reference to the drawings, which should not be construed as limiting the present application. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. It will be apparent, however, to one ordinarily skilled in the art that the specific details need not be employed to practice the present application. In other instances, well-known structures, materials, or processes have not been described in detail in order to avoid obscuring the present application.

[0053] Reference throughout this specification to "one embodiment", "an embodiment", "one example", or "an example" means that a particular feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment of the application. Thus, appearances of the phrases "in one embodiment" or "in an embodiment" or "one example" or "an example" in various places throughout this specification are not necessarily all referring to the same embodiment or example. Furthermore, the particular features, structures, or characteristics can be combined in any suitable

[0054] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The use of the terms "including," "comprising," or "having" and the like herein is specifically intended to convey that although items, components, or elements can not be present, the terminology in some cases conveys the possibility that such items, components, or elements can be present and is thus the possiblity of being present is at least contemplates.

[0055] All terms used herein including technical and scientific terms have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of this specification, and not be interpreted in an idealized or overly formal way.

[0056] In the following, embodiments of the present disclosure will be described in detail with CT scan detection as an example. It should be understood that the embodiments of the present disclosure are not limited to the CT scan detection scenario, and can be applied to various scan detection scenarios. For example, it can be applied to scan detection scenarios including various different inspection objects, including but not limited to vehicle scan detection, luggage / parcel scan detection, human body scan detection, etc. It should be noted that the description of the scan detection scenario here is not an exhaustive description, and the description here should not be understood as limiting the protection scope of the present disclosure.

[0057] It should be noted that the embodiments of the present disclosure are not limited to the target detection task, and can also be applied to target classification, image segmentation, etc. Thus, in this paper, unless otherwise stated, the expression "target recognition" includes but is not limited to various tasks such as target detection, target classification, image segmentation, etc. It should be noted that the embodiments of the present disclosure are not limited to the target detection task, and can also be applied to target classification, image segmentation, etc. Thus, in this paper, unless otherwise stated, the expression "target recognition" includes but is not limited to various tasks such as target detection, target classification, image segmentation, etc.

[0058] It is also noted that, in this document, the expression "target" can include, but is not limited to, for example, a target object of interest, an item of interest, etc. in a scanning detection scene.

[0059] Fig. 1 is a schematic diagram of a structure of a ray scanning detection system according to some exemplary embodiments of the present application. In Fig. 1, a CT scanning apparatus is schematically shown as an example of the ray scanning detection system. As shown in Fig. 1, the CT scanning apparatus according to the present embodiment includes a gantry 20, a carrying mechanism 40, a controller 50, a data processing apparatus 60 (e.g., a computer), etc. The gantry 20 includes a ray source 10, such as an X-ray machine, for emitting X-rays for inspection, and a detection and acquisition apparatus 30. The carrying mechanism 40 carries a detection object (e.g., inspected luggage) 70 to pass through a scanning region between the ray source 10 and the detection and acquisition apparatus 30 of the gantry 20, while the gantry 20 rotates around a forward direction of the detection object 70, so that the rays emitted by the ray source 10 can penetrate the detection object 70, and the detection object 70 is subjected to CT scanning. The detection and acquisition apparatus 30, for example, includes a detector and a data acquisition device having an integral module structure, such as a flat panel detector, for detecting the rays penetrating the detection object 70 (e.g., inspected luggage), obtaining analog signals, and converting the analog signals into digital signals, so as to output projection data of the detection object 70 against X-rays. The controller 50 is used to control the synchronous operation of various parts of the whole system. The data processing apparatus 60 is used to process the data acquired by the data acquisition device, process and reconstruct the data, and output results.

[0060] As shown in Fig. 1, the ray source 10 is placed on one side where the detection object can be placed, and the detection and acquisition apparatus 30 is placed on the other side of the detection object 70, including a detector and a data acquisition device, for obtaining transmission data and / or multi-angle projection data of the detection object 70. The data acquisition device includes a data amplification and shaping circuit therein, which can work in an (electric current) integration mode or a pulse (counting) mode. A data output cable of the detection and acquisition apparatus 30 is connected to the controller 50 and the data processing apparatus 60, and the acquired data is stored in the data processing apparatus 60 according to a trigger command.

[0061] In embodiments of the present disclosure, a ray scanning detection module can be deployed in the data processing device 60, and an image recognition model or a target recognition model can be deployed in the ray scanning detection module. The ray scanning detection module can utilize the image recognition model or the target recognition model to recognize the collected field data (e.g., a ray scanning image), and the recognition can include target detection, e.g., detecting a target object or an object of interest in the ray scanning image. For example, the target object or the object of interest can be various contraband. It should be understood that the specific category of the target object or the object of interest depends on the scanning detection field where the ray scanning detection system is arranged, or in other words, the specific category of the target object or the object of interest is determined by the user of the ray scanning detection system (also referred to as a user herein) according to the specific scanning detection needs, which can be dynamically adjusted according to the scanning detection needs.

[0062] It should be noted that, for the convenience of description, the target object or the object of interest herein can be collectively referred to as a target.

[0063] In embodiments of the present disclosure, the ray source 10 can be an X-ray machine, for example, and a suitable X-ray machine focal point size can be selected according to the imaging resolution. In other embodiments, an X-ray machine can not be used, but a linear accelerator or the like can be used to generate an X-ray beam.

[0064] The detection and acquisition device 30 includes an X-ray detector and a data acquisition circuit, etc. The X-ray detector can use a solid detector, a gas detector or other detectors, and embodiments of the present disclosure are not limited thereto. The data acquisition circuit includes a readout circuit, an acquisition trigger circuit and a data transmission circuit, etc.

[0065] The combination of the controller 50 and the data processing device 60, for example, includes a computer device installed with a control program and a data processing program, responsible for completing the control of the operation process of the CT scanning device, including mechanical rotation, electrical control, safety interlock control, etc., reconstructing a CT image from projection data, training an image recognition model or a target recognition model, and recognizing a ray scanning image using the trained image recognition model or target recognition model, etc.

[0066] FIG. 2 is a schematic diagram of a CT scanning device implementing a reconstruction scanning mode according to some exemplary embodiments of the present disclosure. As shown in FIG. 2, a portion of the detection object 70 in the X-ray transmission field of view 45 emitted from the ray source 10 is received by the detection and acquisition device 30, converted into an electrical signal and further converted into a digital signal representing an attenuation value, as projection data, reconstructed by a computer to obtain a high-quality image.

[0067] According to an embodiment of the present disclosure, the CT scanning device is used to perform CT scanning on the detection object 70 to obtain raw attenuation signals. The attenuation signal data corresponding to a row of detectors can also be displayed in the form of a two-dimensional image. FIG. 3 shows an example of attenuation signal data obtained according to an embodiment of the present disclosure. As shown in FIG. 3, the horizontal axis direction of the raw attenuation signal represents a sequence of detector column directions (for example, from 1 to 256), and the vertical axis represents an angle (for example, from 1 degree to 360 degrees). The raw attenuation signal is pre-processed to become projection data. For example, the projection data can be pre-processed by the CT scanning device to obtain the projection data by performing a negative logarithmic transformation and the like. Then, the data processing device 60 executes a reconstruction program, for example, a mainstream analytical reconstruction method in the related technical field can be used to obtain a reconstructed image, which is the ray scanning image mentioned in the embodiments of the present disclosure.

[0068] In an embodiment of the present disclosure, the data processing device 60 can perform target recognition (for example, target detection) on the above-mentioned reconstructed image (that is, the ray scanning image) to identify the target (for example, contraband) in the ray scanning image. Specifically, a ray scanning detection module can be deployed in the data processing device 60, and an image recognition model or a target recognition model can be deployed in the ray scanning detection module. The ray scanning detection module can use the image recognition model or the target recognition model to recognize the ray scanning image.

[0069] The applicant has found that three-dimensional CT data usually contains a large number of voxels, has sparsity, and also contains certain noise and redundant information. Identifying a three-dimensional CT target will face the problem of huge amount of calculation. In the security check scene, it is difficult to identify a target with strong three-dimensional shape features but complex material composition and physical properties. At the same time, the real-time performance and accuracy of the three-dimensional CT data target recognition in the security check scene are also required to be higher.

[0070] In view of the above problems, the applicant proposes a method for identifying sheet and block contraband, explosives or drugs (see the published patent of the same applicant as the present disclosure, the authorized publication number of which is CN103971084B, and the invention name is “Method for identifying sheet and block contraband, explosives or drugs”, hereinafter, this patent can be referred to as patent document 1). In the patent document 1, the complex three-dimensional target extraction problem is simplified into a two-dimensional target extraction problem which is more mature in technology. The length, thickness and other geometric properties of the two-dimensional area of sheet and block contraband are directly measured by using layer-by-layer tomographic analysis, and the suspicious object is accurately and quickly located. Then, the tomographic area is gradually increased to grow into a three-dimensional object and the physical properties are judged. Finally, the three-dimensional recognition result of the target is obtained, and the effective identification of sheet and block contraband, explosives or drugs is realized.

[0071] The applicant has further found that in the above method, using a limited number of two-dimensional views for target recognition can result in insufficient expression of three-dimensional features, and can lose part of the three-dimensional shape features, spatial texture information, etc. At the same time, since real-time tomographic data analysis is used, different tomographic data are independent of each other during analysis, and the information of tomographic data at a later spatial position cannot affect the analysis results of tomographic data at an earlier position, resulting in serious information loss. If the three-dimensional object grown from the tomographic analysis in the method of Patent Literature 1 is used in combination with comprehensive and three-dimensional CT original data to further identify and judge the suspect, the prohibited and safe items can be further distinguished, and the recognition accuracy can be improved.

[0072] Therefore, an embodiment of the present disclosure proposes a cascaded target recognition method for CT data, which can effectively utilize the feature expression of a three-dimensional object and further improve the recognition accuracy. In the method provided by the embodiment of the present disclosure, a preliminary recognition is first performed on three-dimensional CT data to obtain a first object description set of a target, and then a refined recognition method is cascaded, that is, the first object description set of the target obtained is taken as an input to further recognize a three-dimensional region of interest. In addition, if the three-dimensional region of interest of the target is not included in the pre-obtained three-dimensional recognition result, the method provided by the embodiment of the present disclosure can also be used to process, and the whole or part of the three-dimensional CT data can be taken as an input of the cascaded refined recognition method for recognition.

[0073] For example, some exemplary embodiments of the present disclosure provide a cascaded CT data target recognition method, which includes: obtaining three-dimensional CT data; performing preliminary recognition on the three-dimensional CT data to obtain a first object description set of a target; and performing refined recognition on the target by using a cascaded refined recognition method according to the first object description set of the target to obtain a refined recognition result of the target, wherein in the cascaded refined recognition method, at least one of a first deep learning network for three-dimensional CT data and a second deep learning network for two-dimensional data is used to perform refined recognition on the target to obtain the refined recognition result of the target. For example, in the cascaded refined recognition method, the first deep learning network for three-dimensional CT data can be used to perform refined recognition on the target to obtain the refined recognition result of the target. For example, in the cascaded refined recognition method, the second deep learning network for two-dimensional data can be used to perform refined recognition on the target to obtain the refined recognition result of the target. For example, in the cascaded refined recognition method, the first deep learning network for three-dimensional CT data and the second deep learning network for two-dimensional data can be used to perform refined recognition on the target to obtain the refined recognition result of the target.

[0074] In this embodiment, after obtaining the first object description set of the target by using the primary identification method, the cascaded fine identification method is used for further identification, the three-dimensional information of the CT data is effectively utilized, the prohibited articles and safe articles can be further distinguished, and the target identification accuracy is improved.

[0075] FIG. 4A is a schematic flowchart of a cascaded CT data target identification method according to some example embodiments of the present disclosure. Referring to FIG. 4A, the cascaded CT data target identification method can include the following operations or steps S410-S430.

[0076] Referring to FIG. 4A, in operation or step S410, three-dimensional CT data is acquired.

[0077] With reference to FIGS. 1-3, a portion of the detection object 70 in the X-ray transmission field of view 45 emitted from the ray source 10 is received by the detection and acquisition device 30, converted into an electrical signal and further converted into a digital signal representing an attenuation value as projection data; then, a three-dimensional reconstruction image can be obtained by using a mainstream analytical reconstruction method in the related technical field.

[0078] In an embodiment of the present disclosure, the three-dimensional CT data can include the three-dimensional reconstruction image.

[0079] Referring to FIG. 4A, in operation or step S420, primary identification is performed on the three-dimensional CT data to obtain a first object description set of a target, wherein the number of targets can be N, and N is a positive integer greater than or equal to 1.

[0080] FIG. 5 is a schematic flowchart of a primary identification operation in a cascaded CT data target identification method according to an embodiment of the present disclosure. Referring to FIG. 5, the primary identification operation can include the following sub-operations or sub-steps S510-S540.

[0081] Referring to FIG. 5, in sub-operation or sub-step S510, a slice region extraction step is performed. Specifically, data of a tomographic section of the detection object is acquired; the tomographic section data is preprocessed; the preprocessed tomographic section data is segmented into a plurality of regions with similar physical properties; and in response to a certain segmented region being a slice region, the slice region is extracted.

[0082] Optionally, in sub-operation or sub-step S510, a block region extraction step is performed. Specifically, data of a tomographic section of the detection object is acquired; the tomographic section data is preprocessed; the preprocessed tomographic section data is segmented into a plurality of regions with similar physical properties; and in response to a certain segmented region being a block region, the block region is extracted.

[0083] In the sub-operation or sub-step S520, a slice object forming step is performed. Specifically, it is determined whether the slice region extracted in the current slice can be merged with the slice region extracted in the previous slice; and in response to the slice region extracted in the current slice being able to be merged with the slice region extracted in the previous slice, the slice object is formed by merging.

[0084] Optionally, in the sub-operation or sub-step S520, a block object forming step is performed. Specifically, it is determined whether the block region extracted in the current slice can be merged with the block region extracted in the previous slice; and in response to the block region extracted in the current slice being able to be merged with the block region extracted in the previous slice, the block object is formed by merging.

[0085] In the sub-operation or sub-step S530, it is determined whether the current slice is the last slice.

[0086] In the sub-operation or sub-step S540, in response to the current slice not being the last slice, the slice region extracting step and the slice object forming step are repeatedly performed, and the data of each slice is processed by using a slice-by-slice analysis method.

[0087] Optionally, in the sub-operation or sub-step S540, in response to the current slice not being the last slice, the block region extracting step and the block object forming step are repeatedly performed, and the data of each slice is processed by using a slice-by-slice analysis method.

[0088] That is, in the primary identification operation, each slice data is extracted and processed layer by layer until all the slice data is processed, so as to obtain the first object description set of the target.

[0089] It should be noted that the specific implementation of the primary identification operation (i.e., the operation S420 described above) or step in the target identification method of the cascaded CT data according to the embodiments of the present disclosure can refer to the method described in the patent document 1, that is, the method described in the patent document 1 is wholly or partially incorporated herein as the primary identification method in the embodiments of the present disclosure without conflict.

[0090] It should be further noted that the primary identification operation (i.e., the operation S420 described above) or step in the target identification method of the cascaded CT data according to the embodiments of the present disclosure is not limited to the above implementation.

[0091] Exemplarily, the operation S420 described above can include: processing the three-dimensional CT data using a three-dimensional segmentation network based on deep learning to obtain the first object description set of the target.

[0092] Exemplarily, the operation S420 can include processing the three-dimensional CT data by using a three-dimensional target detection network to obtain the first object description set of the target.

[0093] Exemplarily, the operation S420 can include processing the three-dimensional CT data by using a target feature library comparison manner to obtain the first object description set of the target.

[0094] In embodiments of the present disclosure, after the preliminary identification operation, the obtained first object description set of the target can include one or more of three-dimensional voxel data, point cloud data, feature vectors and semantic descriptions. The semantic descriptions include at least one of category information, position information and confidence of the target.

[0095] Referring to FIG. 4A, in operation or step S430, according to the first object description set of the target, a cascaded fine identification method is used to finely identify the target to obtain a fine identification result of the target.

[0096] In embodiments of the present disclosure, in the cascaded fine identification method, at least one of a first deep learning network for three-dimensional CT data and a second deep learning network for two-dimensional data can be selectively used to finely identify the target to obtain the fine identification result of the target.

[0097] In embodiments of the present disclosure, according to at least one factor in the first object description set of the target and the specific application scene (for example, a specific security check scene) of the method, the first deep learning network and / or the second deep learning network can be selectively configured to perform fine identification.

[0098] For example, in some embodiments of the present disclosure, according to the characteristics of the first object description set of the target, in the cascaded fine identification method, the first deep learning network for three-dimensional CT data can be selectively used to finely identify the target to obtain the fine identification result of the target. That is, in operation or step S430, the first object description set can be input into the first deep learning network for three-dimensional CT data, and the first deep learning network is used to process the first object description set to obtain the fine identification result of the target.

[0099] For example, in some embodiments of the present disclosure, according to the characteristics of the first object description set of the target, in the cascaded fine identification method, the second deep learning network for two-dimensional data can be selectively used to finely identify the target to obtain the fine identification result of the target.

[0100] For example, in some embodiments of the present disclosure, according to characteristics of the first object description set of the target, in the cascaded fine recognition method, both the first deep learning network for three-dimensional CT data and the second deep learning network for two-dimensional data can be selectively utilized to finely recognize the target to obtain a fine recognition result of the target.

[0101] FIG. 4B is a schematic flowchart of a cascaded CT data target recognition method according to some other exemplary embodiments of the present disclosure. Referring to FIG. 4B, the cascaded CT data target recognition method can further include the following operations or steps S4301-S4302.

[0102] In operation or step S4301, a second object description set is generated according to the first object description set of the target, wherein the first object description set is a first object description set for three-dimensional data, and the second object description set is a first object description set for two-dimensional data.

[0103] In operation or step S4302, a fine recognition method is utilized to finely recognize the target to obtain a fine recognition result of the target.

[0104] For example, in operation or step S4302, the second object description set can be input into the second deep learning network for two-dimensional data, and the second object description set is processed by the second deep learning network to obtain the fine recognition result of the target.

[0105] For example, in operation or step S4302, the first object description set can be input into the first deep learning network for three-dimensional CT data, and the first object description set is processed by the first deep learning network; and the second object description set can be input into the second deep learning network for two-dimensional data, and the second object description set is processed by the second deep learning network.

[0106] In some exemplary embodiments, the second object description set includes one or more of slice data of a three-dimensional region of interest of the target along any direction, a feature vector corresponding to the target in the slice data, and a semantic description.

[0107] For example, the three-dimensional region of interest can include a three-dimensional suspect region. Accordingly, the second object description set can include a description set of all or part of the slice data of the three-dimensional suspect region in any direction and its feature vector, semantic description, etc. Specifically, the three-dimensional data can be sliced in any direction at any resolution to obtain a complete slice of the three-dimensional data in that direction. For ease of description, the complete slice can be denoted as a first data form. Then, region extraction can be performed on the slice to obtain all or part of the data set of the slice data. The all or part of the data set of the slice data can be denoted as a second data form. The finally presented slice data form can be the first data form or the second data form alone, or a combination of the first data form and the second data form.

[0108] In some exemplary embodiments, the second object description set includes one or more of a two-dimensional projection image obtained by projecting the three-dimensional region of interest of the target, a feature vector corresponding to the target in the two-dimensional projection image, and a semantic description.

[0109] For example, the three-dimensional suspect region can be projected to obtain one or more two-dimensional images, and the feature vector and semantic description corresponding to the suspect object can be projected into the two-dimensional images in a corresponding method to obtain a two-dimensional projection image.

[0110] In some exemplary embodiments, the second object description set includes one or more of a two-dimensional projection image obtained by projecting the three-dimensional CT data as a whole, a feature vector corresponding to the target in the two-dimensional projection image, and a semantic description.

[0111] For example, the three-dimensional data as a whole can be projected to obtain one or more two-dimensional images, and the three-dimensional data, feature vector, and semantic description corresponding to the suspect object can be mapped into the two-dimensional images. The two-dimensional image data and the mapping obtained suspect object description information can be used as the second object description set and as the input of the fine recognition.

[0112] In some exemplary embodiments, the three-dimensional CT data includes a three-dimensional image obtained by a first CT detection device. The second object description set includes one or more of a two-dimensional image obtained by the first CT detection device or another detection device different from the first CT detection device, a feature vector corresponding to the target in the two-dimensional image, and a semantic description. For example, the other detection device can include a DR detection device.

[0113] It should be noted that in the embodiment of the present disclosure, in the cascaded fine identification method, the input can include one or more of the following: (1) three-dimensional suspect region data corresponding to the suspect object and its feature vector, semantic description, and other description set, where the three-dimensional suspect region data corresponding to the suspect object refers to the local region in the original three-dimensional CT data corresponding to the body data or point cloud data; (2) background region information in the three-dimensional CT image data; (3) other three-dimensional data corresponding to the original three-dimensional image obtained by the original device hardware, and the three-dimensional data corresponding to the suspect object, the feature vector, and the semantic description are mapped to the new three-dimensional image in a predetermined manner, thereby obtaining the three-dimensional image; (4) one or more of the three-dimensional slice data of the target region of interest in any direction, the feature vector corresponding to the target in the slice data, and the semantic description; (5) one or more of the two-dimensional projection image obtained by projecting the three-dimensional region of interest of the target, the feature vector corresponding to the target in the two-dimensional projection image, and the semantic description; (6) one or more of the two-dimensional projection image obtained by projecting the entire three-dimensional CT data, the feature vector corresponding to the target in the two-dimensional projection image, and the semantic description; (7) two-dimensional image obtained by the first CT detection device or other detection device different from the first CT detection device, one or more of the feature vector corresponding to the target in the two-dimensional image, and the semantic description.

[0114] It should be noted that for the above input (3), it can be understood that: the three-dimensional CT data includes a first three-dimensional image obtained by a first CT detection device; the input of the first deep learning network further includes second three-dimensional data obtained by the first CT detection device or other detection device different from the first CT detection device, wherein the second three-dimensional data includes a second three-dimensional image, the second three-dimensional image corresponds to the first three-dimensional image, and the information in the first object description set of the target is mapped to the second three-dimensional image. The "other detection device" here can include a second CT detection device, or other types of detection devices that can obtain three-dimensional images.

[0115] In some exemplary embodiments, in the cascaded fine identification method, the first deep learning network for three-dimensional CT data and the second deep learning network for two-dimensional data are used to finely identify the target to obtain a fine identification result of the target.

[0116] Exemplarily, all or at least part of the first object description set is taken as the input of the first deep learning network, and all or at least part of the second object description set is taken as the input of the second deep learning network. For example, the input of the first deep learning network can include at least one of the above inputs (1), (2), (3), and the input of the second deep learning network can include at least one of the above inputs (2), (4), (5), (6), and (7).

[0117] FIG. 6 is a schematic flowchart of a fine recognition operation in a cascaded CT data target recognition method according to an embodiment of the present disclosure. Exemplarily, referring to FIG. 6, the fine recognition operation can include the following sub-operations or sub-steps S610-S640. It should be noted that in FIG. 6, some of the sub-operations or sub-steps S610-S640 are optional operations or steps.

[0118] In some exemplary embodiments of the present disclosure, in the sub-operation or sub-step S620, feature extraction is performed by using a deep learning network pre-trained for three-dimensional CT data (which can be referred to as a first deep learning network herein) and / or by using a deep learning network pre-trained for two-dimensional data (which can be referred to as a second deep learning network herein) to obtain a feature map. That is, the cascaded three-dimensional target recognition method can include a three-dimensional target recognition method based on a deep learning network and / or a two-dimensional target recognition method based on a deep learning network.

[0119] In embodiments of the present disclosure, the cascaded three-dimensional target recognition method can recognize the three-dimensional region of interest, support input of various three-dimensional data such as single-channel, multi-channel, etc., and be compatible with various information including original three-dimensional CT data, the three-dimensional image semantic description set, etc.

[0120] Optionally, in the sub-operation or sub-step S610, before the operation of feature extraction, the result of the preliminary recognition (for example, the result of the preliminary recognition includes the three-dimensional region of interest) can be pre-processed. The pre-processing method includes but is not limited to operations such as denoising, morphological transformation, geometric transformation, image type conversion, etc., and supports conversion of three-dimensional volume data into point cloud data. For example, the pre-processing can include format conversion, size adjustment, etc. For another example, the three-dimensional region of interest can be pre-processed by rotation correction, denoising, etc. to ensure image quality.

[0121] Rotation correction can help correct the tilt of the image. For example, rotation correction of at least one of the first image, the second image, and the third image can be performed using an image edge contour-based rotation correction method. Specifically, the image edge contour-based rotation correction method can include the following steps: image graying: converting a color image to a grayscale image for subsequent processing; threshold binarization: converting the grayscale image to a binary image so that the target and background in the image have a clear distinction; detecting the contour: detecting the edge contour of the target in the binary image; extracting the enclosing matrix of the contour: finding the smallest rectangular region enclosing the target contour, the edges of this rectangular region can be used to estimate the tilt angle of the image; obtaining the deflection angle through the extracted enclosing matrix: calculating the angle between the long side of the enclosing matrix and the horizontal line, which is the tilt angle of the image; using affine transformation to deflect the image: according to the calculated tilt angle, using affine transformation to rotate the image to the normal angle.

[0122] For example, rotation correction of at least one of the first image, the second image, and the third image can be performed using a Fourier transform and Hough line detection-based rotation correction method. Specifically, the Hough line detection-based rotation correction method can include the following steps: image preprocessing and edge detection: first, pre-process the image, such as graying and binarization, then use an edge detection algorithm (such as the Sobel operator) to detect the edges in the image; Hough transform: Hough transform is a technique for detecting straight lines and curves in an image. Here, it is used to detect straight lines in the image. The basic principle of Hough transform is to use the duality of points and lines to convert the lines in the image space to aggregated points in the parameter space, so as to detect whether there is a curve with a given property in the given image; calculate the tilt angle: through Hough transform, the polar coordinate representation (ρ, θ) of each straight line in the image can be obtained. Where, θ represents the tilt angle of the straight line. The tilt angle of the entire image can be obtained by statistically analyzing the tilt angles of all detected straight lines and then taking the average value; image rotation: according to the calculated tilt angle, the image can be rotated by an image rotation algorithm (such as affine transformation) to correct it; output the corrected image.

[0123] In some exemplary embodiments, the three-dimensional region of interest can be denoised using a filter-based method, a model-based method, or a learning-based method. Denoising is an operation to reduce noise interference in an image or data. In three-dimensional region of interest preprocessing, the goal of denoising is to eliminate or reduce irrelevant information introduced by the acquisition process, sensor noise, or other interference factors. For example, denoising methods can include median filtering, Gaussian filtering, mean filtering, etc., which can smooth the data and remove noise points to improve the accuracy and stability of subsequent processing.

[0124] In some exemplary embodiments, the three-dimensional region of interest can be pre-processed using morphological transformations. Morphological transformations are a class of operations based on the shape and structure of images, used to change the geometric structure of images or extract shape features of interest. In three-dimensional region of interest pre-processing, morphological transformations can be used to remove small spots or isolated noise, fill holes, connect adjacent regions, etc. For example, morphological transformations include dilation, erosion, opening operation, and closing operation, etc.

[0125] In some exemplary embodiments, the three-dimensional region of interest can be pre-processed using geometric transformations. Geometric transformations refer to operations on the position, scale, rotation, or affine transformation of objects in three-dimensional space, in order to better match or adjust images or data. In three-dimensional region of interest pre-processing, geometric transformations can be used to correct or align data to conform to a specific reference coordinate system or geometric shape. For example, translation, rotation, scaling, mirroring, etc. transformation operations can be performed.

[0126] In some exemplary embodiments, the three-dimensional region of interest can be pre-processed using image type conversion. For example, image type conversion can convert the original three-dimensional volume data to point cloud data. In three-dimensional region of interest pre-processing, such conversion can help to convert volume data representation into a series of discrete three-dimensional point coordinates for subsequent point cloud processing and analysis. For example, image type conversion methods can include voxelization, surface reconstruction, and sampling, etc.

[0127] In the cascade fine recognition method, in the case where the input includes "background region information in three-dimensional CT image data", the background region information in three-dimensional CT image data can be pre-processed using traditional image processing or analysis statistical methods to obtain related analysis results of the background region. For example, traditional image processing methods include but are not limited to edge detection, texture analysis, feature descriptor, analysis statistical method, etc. The obtained background region analysis results include feature descriptors, value distribution and statistical information, etc. The above analysis results can be directly used as part of the fine recognition input, or can be used as a screening or enhancement condition for fine recognition input. For example, (1) the feature descriptor is used as part of the fine recognition input, which is input into the fine recognition together with other suspect object descriptions; (2) the value distribution and statistical information are used as screening conditions for fine recognition input, and part of the suspect objects are screened out as fine recognition input according to the above information and preset conditions.

[0128] Through the pre-processing method, the data quality can be improved, the noise interference can be reduced, the geometric structure can be adjusted, or the three-dimensional volume data can be converted into point cloud data, providing more accurate and reliable input for subsequent three-dimensional recognition and analysis.

[0129] It should be noted that three-dimensional body data refers to data represented in three-dimensional space with voxels as the basic unit. Voxel can be understood as a pixel in three-dimensional space, and each voxel contains attribute or numerical information at a specific location in space. By dividing a three-dimensional object or scene into a series of voxels and assigning attribute values to each voxel, a three-dimensional body data can be constructed. Three-dimensional body data can be represented in the form of a cube or a cuboid, providing detailed information about the internal and external structure of the object.

[0130] Point cloud data is a dataset composed of a set of discrete three-dimensional point coordinates. Each point contains location information in three-dimensional space, and can also be accompanied by other attributes such as color, normal vector or intensity. Point cloud data can be obtained in different ways, such as laser radar scanning, stereo vision or three-dimensional reconstruction technology. Point cloud data can represent the surface geometry of a three-dimensional object or the distribution of scattered points in a scene, providing a description of the object's shape, surface details and spatial position. Point cloud data has wide applications in computer vision, computer graphics, robotics and other fields, such as target detection and recognition, three-dimensional reconstruction, environmental perception, etc.

[0131] In summary, three-dimensional body data is three-dimensional data with volume information represented by voxels, while point cloud data is a dataset composed of discrete three-dimensional point coordinates, used to represent the surface geometry of an object or the distribution of scattered points in a scene.

[0132] It should be noted that there are some differences between three-dimensional body data and point cloud data in terms of representation and data structure: (1) representation: three-dimensional body data uses voxels as the basic unit, dividing three-dimensional space into regular cubes or cuboids, and each voxel contains attribute or numerical information at a specific location in space. Point cloud data is a dataset composed of a set of discrete three-dimensional point coordinates, each point containing location information in three-dimensional space. (2) Data structure: three-dimensional body data is usually stored in the form of voxel grid, using multi-dimensional array or voxel tree data structure to represent, where each voxel attribute value can represent the density, color, material, etc. of the object. Point cloud data is stored in the form of a set of points, each point can contain additional attributes such as color, normal vector or intensity. (3) Data density: since three-dimensional body data uses regular voxel grid representation, it can provide detailed information about the internal and external structure of the object, including density distribution, shape, etc. Point cloud data is a set of discrete point coordinates, usually only providing geometric information of the object surface, and the density and distribution of points depend on the method and parameters of collecting or generating the point cloud. (4) Data size: since point cloud data only stores discrete point coordinates, the storage space is more compact compared to three-dimensional body data. Therefore, for large-scale three-dimensional scenes or complex objects, point cloud data may be more suitable for storage and processing.

[0133] Referring to FIG. 6, in sub-operation or sub-step S620, the pre-processed three-dimensional data is subjected to feature extraction by using a first deep learning network for three-dimensional CT data, specifically, by using a specially designed feature extraction network; and / or the pre-processed two-dimensional data is subjected to feature extraction by using a second deep learning network for two-dimensional data, specifically, by using a specially designed feature extraction network.

[0134] In embodiments of the present disclosure, the extracted features include, but are not limited to, shape, texture, color, etc. of the three-dimensional data. Among them, the shape of the three-dimensional data corresponds to the large structure of the target, the texture corresponds to the small structure or detailed structure of the target, and the color corresponds to the z value, density, etc. of the physical properties of the target.

[0135] In embodiments of the present disclosure, useful feature representations are extracted from input data by neural networks or other methods. These features can include shape, texture, color, etc. of the object, which are used to describe and express the key characteristics of the input data. That is, in embodiments of the present disclosure, various information of three-dimensional data can be comprehensively utilized, which is conducive to further improving the accuracy of recognition. In the deep learning network of three-dimensional CT data, by using three-dimensional convolutional neural network (3D CNN) or attention mechanism-based feature extraction method, local and global features in the input data can be learned, and feature maps with distinguishing and expressive ability can be extracted.

[0136] In some exemplary embodiments, in sub-operation or sub-step S620, feature extraction can use three-dimensional convolutional neural network or attention mechanism-based feature extraction method, etc., which can be used alone or in combination, to extract feature maps of three-dimensional CT data; and / or feature extraction can use two-dimensional convolutional neural network or attention mechanism-based feature extraction method, etc., which can be used alone or in combination, to extract feature maps of two-dimensional data.

[0137] If the three-dimensional mask, class information and confidence information of the voxels are contained in the three-dimensional image semantic description set of the target, accurate generation of the feature vector is also supported. That is, in sub-operation or sub-step S620, the three-dimensional mask, class information and confidence contained in the first three-dimensional image semantic description set can be used to generate the feature vector.

[0138] If the two-dimensional mask, class information and confidence information of the pixels are contained in the image semantic description set of the target, accurate generation of the feature vector is also supported. That is, in sub-operation or sub-step S620, the two-dimensional mask, class information and confidence contained in the first image semantic description set can be used to generate the feature vector.

[0139] During the training of the feature extraction network, various data augmentation techniques can be applied to increase the diversity of the data, including but not limited to rotation, translation, scaling, flipping, data fusion, etc. Data augmentation helps improve the robustness and generalization of the network.

[0140] Optionally, in sub-operation or sub-step S630, operations can be performed on the extracted feature maps, including but not limited to feature map fusion, feature map dimension reduction, etc. These operations can be selected and combined according to specific tasks and network architectures. The purpose is to improve the performance and generalization ability of the model through appropriate operations and processing of the feature maps, so as to better adapt to various complex data and application scenarios.

[0141] Specifically, if the network has multiple branches or multiple levels of feature maps, feature map fusion operations can be performed to obtain more comprehensive and multi-scale feature representations. Fusion operations can use feature map weighted summation, feature map splicing, feature pyramid, Transformer-based feature fusion, etc.

[0142] For example, in the feature map weighted summation fusion method, different levels or sources of feature maps are weighted and summed according to certain weights. Each feature map is multiplied by the corresponding weight, and they are added to obtain the fused feature map. The weights can be obtained by learning or manually set.

[0143] For example, in the feature map splicing fusion method, different levels or sources of feature maps are spliced in the channel dimension. For example, if the channel numbers of two feature maps are C1 and C2, the channel number of the spliced feature map is C1+C2. This way can increase the channel dimension of the feature map, so that the network can utilize multiple levels or sources of information at the same time.

[0144] For example, in the feature pyramid fusion method, the feature pyramid is realized by extracting feature maps at different scales and fusing them. Different sizes of convolution kernels or different levels of feature maps are used to capture information at different scales. Then, these feature maps are fused together to obtain multi-scale feature representations.

[0145] For example, in the Transformer-based feature fusion method, Transformer is a neural network model based on self-attention mechanism, which can be used for encoding and decoding sequence data. In feature fusion, the encoder-decoder structure of Transformer can be used to fuse feature maps. The encoder encodes the input feature maps into context-aware feature representations, and the decoder decodes these feature representations into the final fused feature maps.

[0146] It should be noted that these feature fusion methods can be selected and applied according to specific tasks and network structures. Their goal is to improve the representation ability and performance of the network by effectively combining feature information from different sources.

[0147] Specifically, to improve computational efficiency or reduce feature dimension, a feature map can be subjected to a feature map dimension reduction operation that reduces resolution or spatial dimension, for example, the operation can include but is not limited to global average pooling, down-sampling, convolutional dimension reduction (such as 1x1 convolution, depthwise separable convolution, etc.), and the like.

[0148] In an embodiment of the present disclosure, the first deep learning network for three-dimensional CT data can include but is not limited to a three-dimensional convolutional neural network, such as 3D U-Net, V-Net, 3D DenseNet, 3D ResNet, and the like. It should be noted that 3D U-Net is a three-dimensional convolutional neural network based on the U-Net architecture. It has an encoder-decoder structure, extracts and recovers features through stacking convolution, pooling and up-sampling operations, and can effectively capture context information at different scales. V-Net is a volumetric segmentation network based on 3D CNN, specifically for medical image segmentation tasks. It introduces techniques such as residual connection and batch normalization, learns feature representations of images through multiple 3D convolution and deconvolution layers, and generates pixel-level segmentation results. 3D DenseNet is a network structure extended to three dimensions based on 2D DenseNet. It connects the feature maps of each layer with all previous layers through dense connection, promotes the transmission of features and the flow of information, and thus improves the performance and stability of the model. 3D ResNet is a three-dimensional convolutional neural network extended from 2D ResNet. It uses residual connection and deep network ideas, builds the network by stacking multiple residual blocks, effectively solves the problems of gradient vanishing and model degradation, and improves the feature representation ability and learning ability.

[0149] FIG. 7 is an architecture diagram of a cascaded deep learning network used in a target recognition method according to an exemplary embodiment of the present disclosure.

[0150] Referring to FIG. 7, the cascaded deep learning network is a three-dimensional convolutional neural network based on a U-Net architecture. Its structure is composed of an encoder and a decoder, which extracts features through the stacking of convolution, pooling and upsampling operations. The cascaded deep learning network includes an encoder part and a decoder part. The encoder part is a typical convolutional network, which includes repeated application of convolution, followed by a ReLU activation function and a max-pooling operation after each convolution. For example, in this network, each layer in the encoder part can include two 3x3x3 convolutions, each followed by a ReLU layer, and then a 2x2x2 max-pooling layer with a stride of 2 in each direction. The decoder part mainly includes upsampling operations and convolution operations. The upsampling operation is used to restore the resolution of the image, while the convolution operation is used to combine the information to generate an accurate output. In the decoder part, each layer contains a 2x2x2 deconvolution layer with a stride of 2, followed by two 3x3x3 convolution layers, each followed by a ReLU layer. The cascaded deep learning network also includes a skip connection for connecting the feature maps of the corresponding layers of the encoder part and the decoder part to provide more rich contextual information. In the output layer of the network, a 1x1x1 convolution layer and an appropriate activation function (such as Sigmoid or Sofimax) are used to generate the final result.

[0151] The cascaded deep learning network is characterized by a symmetrical structure between the encoder and the decoder, which can fuse low-level features with high-level features through the skip connection, thereby improving the accuracy and detail preservation ability of the segmentation. In addition, the depth and width of the network can be adjusted to adapt to different tasks and data sets to improve performance and generalization ability.

[0152] In some exemplary embodiments of the present disclosure, in the sub-operation or sub-step S640, the feature map is applied to a classification, segmentation or detection task to obtain a set of three-dimensional image semantic descriptions. The "feature map" here can be the feature map directly extracted by the sub-operation or sub-step S620, or the feature map after the extraction by the sub-operation or sub-step S620 and the processing by S630 of the sub-operation or sub-step S620.

[0153] Specifically, the feature map can be mapped to a class probability distribution using a multi-layer perceptron or other classifier to implement a classification task. Additionally or alternatively, the feature map can be restored to the original image size using deconvolution, bilinear interpolation, etc., and assigning semantic labels to each voxel to implement voxel-level segmentation. Additionally or alternatively, a target detection head can be used to perform classification and regression operations on the feature map to predict the class and location of the target, generate class information, confidence and location information, etc. to implement a target detection task.

[0154] In embodiments of the present disclosure, the three-dimensional image semantic description set is in units of one or more of voxels, three-dimensional regions of interest, and three-dimensional CT images, and contains: category information and / or confidence; or the three-dimensional image semantic description set is in units of three-dimensional regions of interest and / or three-dimensional CT images, and contains: at least one of category information, confidence, and position information of the target. For example, the position information of the target in the three-dimensional CT image can contain a three-dimensional bounding box.

[0155] It is found through research that the three-dimensional image semantic description set obtained after fine identification is more accurate than the result of initial identification, and is a modification and improvement of the semantic description set of the previous stage. In embodiments of the present disclosure, in the fine identification method, fine identification is performed using a two-dimensional deep learning network, which can improve the efficiency of fine identification without significantly reducing the identification accuracy.

[0156] In embodiments of the present disclosure, the target identification method can directly obtain the final identification conclusion by using the fine identification method once. Specifically, the step of processing the input by at least one of the first deep learning network and the second deep learning network is executed once; and the output of at least one of the first deep learning network and the second deep learning network after being executed once is taken as the fine identification result of the target.

[0157] FIG. 8 is a schematic flowchart of a cascaded CT data target identification method according to some embodiments of the present disclosure.

[0158] In embodiments of the present disclosure, referring to FIG. 8, the target identification method can use the fine identification method iteratively until the final identification conclusion is obtained. Specifically, the step of processing the input by at least one of the first deep learning network and the second deep learning network is executed iteratively; and the output of at least one of the first deep learning network and the second deep learning network after being executed iteratively is taken as the fine identification result of the target.

[0159] In the process of iteratively using the fine identification method, the output of the i th(i is a positive integer greater than or equal to 1) round of fine identification can be taken as the input of the i+1 thround of fine identification, or the output of the i thround of fine identification and the result of initial identification in the operation or step S420 can be taken together as the input of the i+1 thround of fine identification, or the output of the i thround of fine identification can be taken as a condition for screening the input information of the i+1 thround of fine identification.

[0160] It should be noted that in embodiments of the present disclosure, the termination condition of iteratively using the fine identification method can include that the number of iterations reaches a preset number of rounds, and the output of fine identification reaches a preset accuracy.

[0161] Based on the above method, the embodiments of the present disclosure further provide a device for cascaded CT data target recognition. The device will be described in detail below in combination with FIG. 9.

[0162] FIG. 9 is a structural block diagram of a device for cascaded CT data target recognition according to an exemplary embodiment of the present disclosure. Referring to FIG. 9, the device 900 includes an acquisition module 910, a preliminary recognition module 920 and a fine recognition module 930.

[0163] The acquisition module 910 is configured to acquire three-dimensional CT data. In some exemplary embodiments, the acquisition module 910 can be configured to perform the operation S410 and its sub-operations described above, which will not be repeated here.

[0164] The preliminary recognition module 920 is configured to preliminarily recognize the three-dimensional CT data to obtain a first object description set of a target. In some exemplary embodiments, the preliminary recognition module 920 can be configured to perform the operation S420 and its sub-operations described above, which will not be repeated here.

[0165] The fine recognition module 930 is configured to finely recognize the target according to the first object description set of the target by using a cascaded fine recognition method to obtain a fine recognition result of the target. In some exemplary embodiments, the fine recognition module 930 can be configured to perform the operation S430 and its sub-operations described above, which will not be repeated here.

[0166] In the fine recognition method performed by the fine recognition module 930, at least one of a first deep learning network for three-dimensional CT data and a second deep learning network for two-dimensional data is used to finely recognize the target to obtain the fine recognition result of the target.

[0167] In some embodiments, the fine recognition module 930 is configured to perform feature extraction by using a pre-trained first deep learning network for three-dimensional CT data and / or a second deep learning network for two-dimensional data to obtain a feature map. For example, a feature vector can be generated by using the three-dimensional mask, the category information and the confidence contained in the first three-dimensional image semantic description set.

[0168] Exemplarily, the extracted features include at least one of shape, texture and color in the three-dimensional CT data.

[0169] Optionally, the fine recognition module 930 is further configured to operate the feature map by using the first deep learning network and / or the second deep learning network, wherein the operation includes at least one of feature map fusion and feature map dimension reduction.

[0170] Exemplarily, the fine recognition module 930 is further configured to apply the feature map to a classification, segmentation or detection task to obtain a three-dimensional image semantic description set.

[0171] In embodiments of the present disclosure, the three-dimensional image semantic description set is in units of one or more of voxels, three-dimensional regions of interest and three-dimensional CT images, and contains class information and / or confidence; or the three-dimensional image semantic description set is in units of three-dimensional regions of interest and / or three-dimensional CT images, and contains at least one of class information, confidence and position information of the target.

[0172] According to embodiments of the present disclosure, any one or more of the acquisition module 910, the coarse recognition module 920 and the fine recognition module 930 and sub-modules thereof included in the apparatus 900 can be combined in one module for implementation, or any one of them can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules, and implemented in one module. According to embodiments of the present disclosure, at least one of the acquisition module 910, the coarse recognition module 920 and the fine recognition module 930 and sub-modules thereof included in the apparatus 900 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on chip, a system on board, a system on package, an application specific integrated circuit (ASIC), or any other reasonable way of integrating or packaging circuits, etc. hardware or firmware, or in any one of software, hardware and firmware implementation or in any appropriate combination of any of them. Alternatively, at least one of the acquisition module 910, the coarse recognition module 920 and the fine recognition module 930 and sub-modules thereof included in the apparatus 900 can be at least partially implemented as computer program modules that can perform corresponding functions when executed.

[0173] FIG. 10 schematically shows a structural block diagram of an electronic device suitable for implementing the target recognition method according to exemplary embodiments of the present disclosure.

[0174] As shown in FIG. 10, the electronic device 1000 according to an embodiment of the disclosure includes a processor 1001 that can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded into a random access memory (RAM) 1003 from a storage section 1008. The processor 1001 can include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor, and / or a related chipset, and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), and the like. The processor 1001 can also include an on-board memory for cache use. The processor 1001 can include a single processing unit or multiple processing units for executing different actions of the method processes according to embodiments of the disclosure.

[0175] For example, the electronic device can be the data processing device 60 of the ray scanning detection system.

[0176] In the RAM 1003, various programs and data required for the operation of the electronic device 1000 are stored. The processor 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. The processor 1001 performs various operations of the method processes according to embodiments of the disclosure by executing the programs in the ROM 1002 and / or the RAM 1003. Note that the programs can also be stored in one or more memories other than the ROM 1002 and the RAM 1003. The processor 1001 can also perform various operations of the method processes according to embodiments of the disclosure by executing the programs stored in the one or more memories.

[0177] According to embodiments of the disclosure, the electronic device 1000 can further include an input / output (I / O) interface 1005, which is also connected to the bus 1004. The electronic device 1000 can further include one or more of the following components connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, etc.; an output section 1007 including a display such as a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, a modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as necessary. A removable medium 1011 such as a magnetic disk, a magneto-optical disk, a semiconductor memory, etc. is attached to the drive 1010 as necessary, so that a computer program read out therefrom is installed into the storage section 1008 as necessary.

[0178] The present disclosure also provides a computer readable storage medium, which can be included in the device / apparatus / system described in the above embodiments, or can exist separately without being assembled into the device / apparatus / system. The above computer readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of the present disclosure.

[0179] According to an embodiment of the present disclosure, the computer readable storage medium can be a non-volatile computer readable storage medium, which can include, but is not limited to, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any appropriate combination thereof. In the present disclosure, the computer readable storage medium can be any tangible medium that contains or stores a program, which can be used by or in connection with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present disclosure, the computer readable storage medium can include one or more memories, such as the ROM 1002 and / or the RAM 1003 described above, and / or one or more memories other than the ROM 1002 and the RAM 1003.

[0180] Embodiments of the present disclosure also include a computer program product, which includes a computer program containing program codes for executing the methods shown in the flowcharts. When the computer program product is run in a computer system, the program codes are used to make the computer system implement the methods provided by the embodiments of the present disclosure.

[0181] The above functions defined in the system / apparatus of the embodiments of the present disclosure are performed when the computer program is executed by the processor 1001. According to an embodiment of the present disclosure, the above described system, apparatus, module, unit, etc. can be implemented by computer program modules.

[0182] In one embodiment, the computer program can rely on a tangible storage medium, such as an optical storage medium, a magnetic storage medium, etc. In another embodiment, the computer program can also be transmitted, distributed, and downloaded in the form of a signal via a network medium, and be downloaded and installed through the communication part 1009 and / or installed from the detachable medium 1011. The program codes contained in the computer program can be transmitted by any appropriate network medium, including but not limited to wireless, wired, etc., or any appropriate combination thereof.

[0183] In such embodiments, the computer program can be downloaded and installed from the network through the communication section 1009, and / or installed from the removable medium 1011. When the computer program is executed by the processor 1001, the above-described functions defined in the system of the embodiments of the present disclosure are performed. According to the embodiments of the present disclosure, the system, device, apparatus, module, unit, and the like described above can be implemented by computer program modules.

[0184] According to the embodiments of the present disclosure, the program code for executing the computer program provided by the embodiments of the present disclosure can be written in any combination of one or more programming languages, and specifically, these computer programs can be implemented by using high-level procedural and / or object-oriented programming language, and / or assembly / machine language. The programming language includes, but is not limited to, such as Java, C++, python, “C” language or similar programming language. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In the case involving a remote computing device, the remote computing device can be connected to the user computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, connected through the Internet by using an Internet service provider).

[0185] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each block in the flowcharts or block diagrams can represent a module, a program segment, or a part of code, which contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur in different order than that shown in the accompanying drawings. For example, two blocks that are shown in succession can actually be executed substantially in parallel, and sometimes they can be executed in reverse order, depending on the involved functions. It should also be noted that each block in the block diagrams or flowcharts, and the combination of blocks in the block diagrams or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0186] The target identification method and device, electronic device, ray scanning detection system, computer readable storage medium and program product provided by the embodiments of the present disclosure have at least one aspect of the following beneficial effects:

[0187] After obtaining the three-dimensional CT data, first, the initial identification is performed, and then the three-dimensional target identification method is cascaded, supporting the result obtained through the initial identification as the input of the cascaded fine identification method. Through the three-dimensional target identification method of cascading the initial identification and the fine identification, the feature expression of the three-dimensional object can be enhanced, and the recognition accuracy can be further improved.

[0188] In particular, in the field of security CT, three-dimensional CT data provides more comprehensive, accurate and visualized information, and effectively utilizing the three-dimensional CT data for identifying contraband is conducive to improving the accuracy of the identification algorithm, so that potential threats and hidden items can be more accurately identified. In the method provided in the embodiments of the present disclosure, after obtaining the suspicious first object description set by using the initial identification method, the fine identification method is cascaded, the three-dimensional information of the CT data is effectively utilized, and the contraband and safe items can be further distinguished, and the identification accuracy is improved.

[0189] The above describes the embodiments of the present disclosure. However, these embodiments are only for illustrative purposes, and are not intended to limit the scope of the present disclosure. Although each embodiment is described above separately, this does not mean that the measures in each embodiment cannot be used advantageously in combination. The scope of the present disclosure is defined by the appended claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art can make various substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A method for identifying a target in CT data in a cascade, characterized by The method comprises: acquiring three-dimensional CT data; performing preliminary identification on the three-dimensional CT data to obtain a first object description set of a target; and performing fine identification on the target according to the first object description set of the target by using a cascaded fine identification method to obtain a fine identification result of the target, wherein in the cascaded fine identification method, at least one of a first deep learning network for three-dimensional CT data and a second deep learning network for two-dimensional data is selectively used to perform fine identification on the target to obtain the fine identification result of the target.

2. The method of claim 1, wherein, The first object description set comprises one or more of three-dimensional voxel data, point cloud data, feature vectors and semantic descriptions.

3. The method of claim 2, wherein, The semantic description comprises at least one of category information, position information and confidence of the target.

4. The method of any one of claims 1-3, wherein, The preliminary identification on the three-dimensional CT data to obtain the first object description set of the target specifically comprises: processing the three-dimensional CT data by using a three-dimensional segmentation network based on deep learning to obtain the first object description set of the target; or processing the three-dimensional CT data by using a three-dimensional target detection network to obtain the first object description set of the target; or processing the three-dimensional CT data by using a target feature library comparison method to obtain the first object description set of the target.

5. The method of any one of claims 1-3, wherein, The preliminary identification on the three-dimensional CT data to obtain the first object description set of the target specifically comprises: a slice region extraction step of acquiring data of a slice of an object to be detected; pre-processing the slice data; segmenting the pre-processed slice data into a plurality of regions with similar physical properties; and extracting a slice region in response to the slice region being a slice region; a slice object formation step of judging whether the slice region extracted in the current slice can be merged with the slice region extracted in the previous slice; and forming a slice object in response to the slice region extracted in the current slice being able to be merged with the slice region extracted in the previous slice; and repeatedly performing the slice region extraction step and the slice object formation step to process data of each slice by using a slice-by-slice analysis method until data of all slices are processed to obtain the first object description set of the target, or The preliminary identification on the three-dimensional CT data to obtain the first object description set of the target specifically comprises: a block region extraction step of acquiring data of a slice of a detection object; pre-processing the slice data; segmenting the pre-processed slice data into a plurality of regions with similar physical properties; and extracting a block region in response to the block region being a block region; a block object formation step of judging whether the block region extracted in the current slice can be merged with the block region extracted in the previous slice; and forming a block object in response to the block region extracted in the current slice being able to be merged with the block region extracted in the previous slice; and The block region extraction step and the block object formation step are repeatedly performed, and data of each slice is processed by a slice-by-slice analysis method until data of all slices are processed, so as to obtain a target object description set of the block target.

6. The method of any one of claims 1-5, wherein, According to the first object description set of the target, a cascaded fine recognition method is used to perform fine recognition on the target, so as to obtain a fine recognition result of the target, including: The first object description set is input into the first deep learning network, and the first object description set is processed by using the first deep learning network, so as to obtain the fine recognition result of the target.

7. The method of any one of claims 1-5, wherein, The method further includes: generating a second object description set according to the first object description set of the target, wherein the first object description set is an object description set for three-dimensional data, and the second object description set is an object description set for two-dimensional data.

8. The method of claim 7, wherein, According to the first object description set of the target, a cascaded fine recognition method is used to perform fine recognition on the target, so as to obtain a fine recognition result of the target, including: The second object description set is input into the second deep learning network, and the second object description set is processed by using the second deep learning network, so as to obtain the fine recognition result of the target.

9. The method of claim 7, wherein, In the cascaded fine recognition method, the first deep learning network and the second deep learning network are used to perform fine recognition on the target, so as to obtain the fine recognition result of the target.

10. The method of claim 9, wherein, The first object description set is input into the first deep learning network, and the first object description set is processed by using the first deep learning network, so as to obtain the fine recognition result of the target. The first object description set is input into the first deep learning network, and the first object description set is processed by using the first deep learning network, so as to obtain the fine recognition result of the target. The first object description set is input into the first deep learning network, and the first object description set is processed by using the first deep learning network, so as to obtain the fine recognition result of the target.

11. The method of claim 6 or 10, wherein, The first object description set is input into the first deep learning network, and the first object description set is processed by using the first deep learning network, so as to obtain the fine recognition result of the target. The first object description set is input into the first deep learning network, and the first object description set is processed by using the first deep learning network, so as to obtain the fine recognition result of the target.

12. The method of claim 11, wherein, The input of the first deep learning network further includes: Background region information in three-dimensional CT image data; and / or, The three-dimensional CT data includes first three-dimensional image obtained by a first CT detection device; and the input of the first deep learning network further includes second three-dimensional data obtained by the first CT detection device or other detection device different from the first CT detection device, wherein the second three-dimensional data includes second three-dimensional image, the second three-dimensional image corresponds to the first three-dimensional image, and information in the first object description set of the target is mapped into the second three-dimensional image.

13. The method of any one of claims 7-10, wherein, The second object description set includes one or more of slice data of a three-dimensional region of interest of the target along any direction, feature vector corresponding to the target in the slice data, and semantic description; and / or, The second object description set comprises one or more of a two-dimensional projection image obtained by projecting a three-dimensional region of interest of the target, a feature vector corresponding to the target in the two-dimensional projection image, and a semantic description; and / or, The second object description set comprises one or more of a two-dimensional projection image obtained by projecting the entire three-dimensional CT data, a feature vector corresponding to the target in the two-dimensional projection image, and a semantic description; and / or, The three-dimensional CT data comprises a first three-dimensional image obtained by a first CT detection device; And the second object description set comprises one or more of a two-dimensional image obtained by the first CT detection device or other detection device different from the first CT detection device, a feature vector corresponding to the target in the two-dimensional image, and a semantic description.

14. The method of claim 6 or 10, wherein, The processing of the first object description set by the first deep learning network specifically comprises feature extraction by a pre-trained first deep learning network to obtain a feature map.

15. The method of claim 8 or 10, wherein, The processing of the second object description set by the second deep learning network specifically comprises feature extraction by a pre-trained second deep learning network to obtain a feature map.

16. The method of claim 14 or 15, wherein, The extracted features comprise at least one of shape, texture, and color in the three-dimensional CT data.

17. The method of claim 14, wherein, The processing of the first object description set by the first deep learning network specifically comprises operations on the feature map by the first deep learning network, wherein the operations comprise at least one of feature map fusion and feature map dimension reduction.

18. The method of claim 15, wherein, The processing of the second object description set by the second deep learning network specifically comprises operations on the feature map by the second deep learning network, wherein the operations comprise at least one of feature map fusion and feature map dimension reduction.

19. The method of claim 14 or 17, wherein, The processing of the first object description set by the first deep learning network further specifically comprises applying the feature map to a classification, segmentation, or detection task to obtain a three-dimensional image semantic description set.

20. The method of claim 15 or 18, wherein, The processing of the second object description set by the second deep learning network further specifically comprises applying the feature map to a classification, segmentation, or detection task to obtain a two-dimensional image semantic description set.

21. The method of any one of claims 1-7, 9-11, 14, 17, and 19, wherein, The fine recognition of the target by the cascaded fine recognition method to obtain a fine recognition result of the target specifically comprises: Single execution of the first deep learning network to process the input; and the output after single execution of the first deep learning network is taken as the fine recognition result of the target.

22. The method of any one of claims 1-5, 7-10, 15, 18, and 20, wherein, The fine recognition of the target by the cascaded fine recognition method to obtain a fine recognition result of the target specifically comprises: single execution of the second deep learning network to process the input; and the output after single execution of the second deep learning network is taken as the fine recognition result of the target.

23. The method of any one of claims 1-7, 9-11, 14, 17, and 19, wherein, The target is finely identified by using a cascaded fine identification method to obtain a fine identification result of the target, specifically including: iteratively executing the step of processing the input by the first deep learning network; and taking the output after the first deep learning network is iteratively executed multiple times as the fine identification result of the target.

24. The method of any one of claims 1-7, 7-10, 15, 18, and 20, wherein, The target is finely identified by using a cascaded fine identification method to obtain a fine identification result of the target, specifically including: iteratively executing the step of processing the input by the second deep learning network; and taking the output after the second deep learning network is iteratively executed multiple times as the fine identification result of the target.

25. The method of any one of claims 1-24, wherein, The fine identification result of the target is in units of one or more of voxels, three-dimensional regions of interest and three-dimensional CT images, and contains: class information and / or confidence; or, The fine identification result of the target is in units of three-dimensional regions of interest and / or three-dimensional CT images, and contains: at least one of class information, confidence and position information of the target.

26. A device for identifying a target in CT data, the device comprising: The apparatus includes: An acquisition module configured to acquire three-dimensional CT data; A preliminary identification module configured to preliminarily identify the three-dimensional CT data to obtain a first object description set of a target; and A fine identification module configured to finely identify the target by using a cascaded fine identification method according to the first object description set of the target to obtain a fine identification result of the target, wherein in the cascaded fine identification method, at least one of a first deep learning network for three-dimensional CT data and a second deep learning network for two-dimensional data is used to finely identify the target to obtain the fine identification result of the target. 27.An electronic device comprising: one or more processors; a memory storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1-25. 28.A radiographic scanning detection system comprising: a radiation source configured to emit radiation; a detector configured to receive radiation emitted from the radiation source and passing through a detection object; and a data processing apparatus in communication connection with the detector, wherein the data processing apparatus comprises: one or more processors; a memory storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1-25. The computer program is executed by a processor to implement the steps of the method according to any one of claims 1-25.

29. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by a processor to implement the steps of the method according to any one of claims 1-25.

30. A computer program product comprising a computer program, characterized in that, The computer program is executed by a processor to implement the steps of the method according to any one of claims 1-25.

Images