An image-assisted sorting method and sorting system for gastroenterology

Abstract

The present invention provides an image-assisted sorting method and a sorting system for gastroenterology, which relate to the technical field of medical image processing, and include acquiring and processing images to be sorted corresponding to the digestive tract; using a first classifier to classify the images to be sorted according to digestive structures, and outputting sub-images corresponding to different digestive structures; based on preset condition one, using a second classifier to perform a lesion detection on all sub-images once to obtain normal areas and lesion areas; based on preset condition two, using a third classifier to perform a secondary lesion detection on the lesion areas to obtain suspected lesion areas and confirmed lesion areas; distinguishing and marking the images to be sorted according to the normal areas, suspected lesion areas and confirmed lesion areas, and visually displaying the marked images to be sorted. The present invention combines the first classifier, the second classifier and the third classifier to perform multi-level classification and lesion detection, and uses visual display to improve the convenience of doctors' diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to an image-assisted sorting method and a sorting system for gastroenterology. Background Art

[0002] Image collation in gastroenterology is an important part of medical diagnosis. With the development of medical imaging technology, a large amount of digestive tract image data has been acquired, but the collation and analysis of these images still face many challenges. Traditional image collation methods mainly rely on manual operations, which are inefficient and prone to errors. Therefore, it is urgent to develop a method that can automatically collate and analyze gastroenterology images.

[0003] At present, although there are some automatic image sorting methods, they still have problems such as low accuracy and slow processing speed in classification and lesion detection. Especially when processing complex digestive tract images, it is often difficult to accurately distinguish normal areas, suspected lesion areas, and confirmed lesion areas.

[0004] Therefore, it is necessary to provide an image-assisted organization method and organization system for gastroenterology to solve the above-mentioned technical problems. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides an image-assisted sorting method and sorting system for gastroenterology, which combines a first classifier, a second classifier and a third classifier to perform multi-level classification and lesion detection, and uses visual display to improve the convenience of doctors' diagnosis.

[0006] In one aspect, the present invention provides an image-assisted sorting method for gastroenterology, the sorting method comprising the following steps:

[0007] Acquiring and processing the image to be sorted corresponding to the digestive tract;

[0008] Using a first classifier to classify the image to be sorted according to the digestive structure, and outputting sub-images corresponding to different digestive structures, wherein the first classifier is trained based on sample images;

[0009] Based on the preset condition 1, the second classifier is used to perform lesion detection on all sub-images to obtain normal areas and lesion areas;

[0010] Based on the second preset condition, a third classifier is used to perform secondary lesion detection on the lesion area to obtain a suspected lesion area and a confirmed lesion area;

[0011] The image to be sorted is distinguished and marked according to the normal area, the suspected lesion area and the confirmed lesion area, and the marked image to be sorted is visually displayed.

[0012] Preferably, the image to be sorted and the sample image are one of an endoscopic image, an X-ray angiography image, a CT image and an MRI image of the digestive tract.

[0013] Preferably, the second classifier and the third classifier each include sub-classification modules corresponding to different digestive structures, wherein the digestive structures include stomach, liver, and small intestine.

[0014] Preferably, the acquiring and processing of the image to be sorted corresponding to the digestive tract includes:

[0015] Acquire an image to be sorted corresponding to the digestive tract;

[0016] The image to be sorted is preprocessed, wherein the preprocessing includes noise reduction, artifact removal, and contrast enhancement.

[0017] Preferably, the training process of the first classifier includes:

[0018] Collecting sample images, wherein each sample image includes a sample digestion structure and a sample feature corresponding to the sample digestion structure;

[0019] Using the sample features of each sample image as input and the sample digestion structure as output to train a first classifier, wherein the first classifier adopts a deep learning framework and is built based on a neural network;

[0020] Output the digestive structure label corresponding to each sample image.

[0021] Preferably, the first classifier is used to classify the image to be sorted according to the digestive structure, and output sub-images including different digestive structures, including:

[0022] Segmenting the image to be sorted based on an edge detection method to obtain a segmented image including different digestive structures;

[0023] Extracting features of each segmented image based on a Gabor filter, wherein the features include shape, size, texture, and boundary curvature;

[0024] Based on the features of the segmented image, the segmented image is identified and classified using a trained first classifier, and sub-images corresponding to different digestive structures are output.

[0025] Preferably, based on the preset condition 1, the second classifier is used to perform lesion detection on all sub-images to obtain normal areas and lesion areas, including:

[0026] Setting a preset condition 1, wherein the preset condition 1 is a first threshold condition;

[0027] All sub-images are input into the second classifier for detection in sequence.

[0028] If the sub-image meets the first threshold condition, it is marked as a diseased area, otherwise, it is marked as a normal area;

[0029] The second classifier outputs a first detection result including a normal area and a lesion area.

[0030] Preferably, based on the second preset condition, using a third classifier to perform secondary lesion detection on the lesion area to obtain a suspected lesion area and a confirmed lesion area includes:

[0031] Setting a second preset condition, wherein the second preset condition is a second threshold condition;

[0032] All lesion areas are input into the third classifier for detection in sequence.

[0033] If the lesion area meets the second threshold condition, it is marked as a confirmed lesion area, otherwise, it is marked as a suspected lesion area;

[0034] The third classifier outputs a second detection result including a suspected lesion area and a confirmed lesion area.

[0035] Preferably, the step of distinguishing and marking the images to be sorted according to the normal areas, the suspected lesion areas and the confirmed lesion areas, and visually displaying the marked images to be sorted includes:

[0036] Color marking is performed on the normal areas, suspected lesion areas and confirmed lesion areas classified in the images to be sorted;

[0037] Perform visualization operations on the images to be sorted after color marking.

[0038] On the other hand, the present invention also provides an image-assisted sorting system for gastroenterology, which can at least be used to implement the above sorting method, including:

[0039] An image processing module, used for acquiring and processing the image to be sorted corresponding to the digestive tract;

[0040] A first classification module, used to classify the image to be sorted according to the digestive structure using a first classifier, and output sub-images corresponding to different digestive structures, wherein the first classifier is trained based on sample images;

[0041] The second classification module is used to perform lesion detection on all sub-images based on the preset condition 1 by using the second classifier to obtain normal areas and lesion areas;

[0042] A third classification module is used to perform secondary lesion detection on the lesion area using a third classifier based on preset condition 2 to obtain a suspected lesion area and a confirmed lesion area;

[0043] The identification module is used to distinguish and identify the images to be sorted according to the normal area, the suspected lesion area and the confirmed lesion area, and to visually display the identified images to be sorted.

[0044] Compared with the related art, the image-assisted sorting method and sorting system for gastroenterology provided by the present invention has the following beneficial effects:

[0045] The present invention combines the first classifier, the second classifier and the third classifier to perform multi-level classification and lesion detection on the image to be sorted, obtains normal areas, suspected lesion areas and confirmed lesion areas, and performs differentiated visual display to improve the convenience of doctors' diagnosis. At the same time, it can realize efficient and accurate analysis of digestive tract images and improve the efficiency and accuracy of medical diagnosis.

[0046] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0048] Figure 1 A flowchart of an image-assisted arrangement method for gastroenterology provided by the present invention;

[0049] Figure 2 A flowchart of step S1 in a gastroenterology image-assisted arrangement method provided by the present invention;

[0050] Figure 3 A flowchart of step S2 in a gastroenterology image-assisted arrangement method provided by the present invention;

[0051] Figure 4 A flowchart of step S3 in a gastroenterology image-assisted arrangement method provided by the present invention;

[0052] Figure 5 A flowchart of step S4 in a gastroenterology image-assisted arrangement method provided by the present invention;

[0053] Figure 6A flowchart of step S5 in a gastroenterology image-assisted arrangement method provided by the present invention;

[0054] Figure 7 This is a module structure diagram of a gastroenterology image-assisted arrangement system provided by the present invention. DETAILED DESCRIPTION

[0055] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0056] It should be noted that the images to be sorted and the sample images can be endoscopic images of the digestive tract, or any one of X-ray contrast images, CT images or MRI images, which are usually used in medical diagnosis and treatment. Endoscopic images can directly observe the inside of the digestive tract, X-ray contrast images can show the outline and function of the digestive tract, CT images can provide structural and position information of the digestive tract, and MRI images can provide more detailed soft tissue information. Embodiment 1

[0057] The present invention provides an image-assisted arrangement method for gastroenterology, referring to Figure 1 As shown, the finishing method comprises the following steps:

[0058] S1: Acquire and process the image to be sorted corresponding to the digestive tract.

[0059] In this embodiment, in order to obtain high-quality image data for subsequent analysis and processing, by acquiring endoscopic images, X-ray angiography images, CT images or MRI images of the digestive tract, a comprehensive understanding of the state and pathological conditions of the digestive tract can be obtained. At the same time, high-quality image data is the basis for accurate medical diagnosis and disease monitoring. Through preprocessing and enhancement, noise and interference can be removed, and the contrast and resolution of the image can be improved, so as to better identify and diagnose lesions in the digestive tract.

[0060] S2: using a first classifier to classify the image to be sorted according to the digestive structure, and outputting sub-images corresponding to different digestive structures, wherein the first classifier is trained based on sample images.

[0061] In this embodiment, the digestive structures in the image to be sorted are automatically classified into different sub-images, thereby facilitating subsequent analysis and processing.

[0062] Among them, the first classifier is trained based on sample images, which means that it has learned the characteristics and patterns of digestive structures from a large number of sample images. Through training, the first classifier can automatically identify and classify the digestive structures in the images to be sorted, and finally output sub-images corresponding to different digestive structures.

[0063] Specifically, the digestive structure includes the stomach, liver, small intestine, etc. Then, multiple sub-images correspond to the stomach, liver, and small intestine. This automatic classification can greatly improve the efficiency and accuracy of medical image processing, which not only simplifies the processing process, but also reduces human errors, providing doctors with more reliable and accurate diagnostic results.

[0064] S3: Based on the preset condition one, the second classifier is used to perform lesion detection on all sub-images to obtain normal areas and lesion areas.

[0065] In this embodiment, this step is mainly to automatically detect lesions in the sub-image and distinguish between normal areas and lesion areas. At the same time, the second classifier is similar to the first classifier and is also trained based on a large number of sample images. The difference is that the second classifier focuses on lesion detection and can learn and identify abnormal areas in the digestive structure. By training the second classifier, the lesion area in the sub-image can be automatically detected and distinguished from the normal area.

[0066] Through automated lesion detection, the lesion area can be located quickly and accurately, improving the accuracy and efficiency of diagnosis. In addition, this step also helps to reduce the workload of doctors, reduce human errors, and provide patients with more reliable medical services.

[0067] S4: Based on the preset condition 2, a third classifier is used to perform secondary lesion detection on the lesion area to obtain a suspected lesion area and a confirmed lesion area.

[0068] In this embodiment, this step is to classify the lesion area more accurately and distinguish between the suspected lesion area and the confirmed lesion area.

[0069] The third classifier is also trained based on a large number of sample images. It focuses on the fine classification of the lesion area. By learning the third classifier, it can identify the characteristics of the lesion area and divide it into suspected lesion areas and confirmed lesion areas based on these characteristics.

[0070] Accurate classification of lesion areas is crucial for medical diagnosis and treatment. A suspected lesion area may have the possibility of lesions, while a confirmed lesion area can be confirmed as a lesion. Fine classification of lesion areas helps doctors to more accurately judge the condition and develop more targeted treatment plans.

[0071] S5: Differentiate and mark the image to be sorted according to the normal area, the suspected lesion area and the confirmed lesion area, and visualize and display the marked image to be sorted.

[0072] In this embodiment, this step is to provide doctors with an intuitive and visual image so that they can quickly and accurately understand and analyze the lesion condition.

[0073] In the process of distinguishing and identifying, different colors or marks are used to mark normal areas, suspected lesion areas and confirmed lesion areas according to their classification results, so that doctors can see the different areas in the image and their corresponding classification results at a glance.

[0074] Visual display presents the annotated images to doctors for observation and analysis. Through visual display, doctors can intuitively observe the location, size and shape of the lesion area, so as to make a more accurate judgment on the condition.

[0075] In addition, the second classifier and the third classifier both include sub-classification modules corresponding to different digestive structures, wherein the digestive structures include the stomach, liver, and small intestine, and the corresponding sub-images are input into the corresponding sub-modules for corresponding lesion detection.

[0076] In this embodiment, if Figure 2 As shown, step S1 specifically includes the following steps:

[0077] S101: Acquire the image to be sorted corresponding to the digestive tract.

[0078] S102: Preprocessing the image to be sorted, wherein the preprocessing includes noise reduction, artifact removal, and contrast enhancement.

[0079] Preferably, Figure 3 As shown, the training process of the first classifier includes:

[0080] S201: Collecting sample images, wherein each sample image includes a sample digestion structure and a sample feature corresponding to the sample digestion structure;

[0081] S202: using the sample features of each sample image as input and the sample digestion structure as output to train a first classifier, wherein the first classifier adopts a deep learning framework and is built based on a neural network;

[0082] S203: Output the digestive structure label corresponding to each sample image.

[0083] Preferably, the first classifier is used to classify the image to be sorted according to the digestive structure, and output sub-images including different digestive structures, including:

[0084] S204: Segmenting the image to be sorted based on an edge detection method to obtain segmented images including different digestive structures;

[0085] S205: extracting features of each segmented image based on a Gabor filter, wherein the features include shape, size, texture, and boundary curvature;

[0086] S206: Based on the features of the segmented image, the segmented image is identified and classified using the trained first classifier, and sub-images corresponding to different digestive structures are output.

[0087] Preferably, Figure 4 As shown, based on the preset condition 1, the second classifier is used to perform lesion detection on all sub-images to obtain normal areas and lesion areas, including:

[0088] S301: Setting a preset condition 1, wherein the preset condition 1 is a first threshold condition;

[0089] S302: input all sub-images into the second classifier in sequence for detection,

[0090] If the sub-image meets the first threshold condition, it is marked as a diseased area, otherwise, it is marked as a normal area;

[0091] S303: The second classifier outputs a first detection result including a normal area and a lesion area.

[0092] Preferably, Figure 5 As shown, based on the preset condition 2, the third classifier is used to perform secondary lesion detection on the lesion area to obtain the suspected lesion area and the confirmed lesion area, including:

[0093] S401: Setting a second preset condition, wherein the second preset condition is a second threshold condition;

[0094] S402: inputting all lesion areas into the third classifier in sequence for detection,

[0095] If the lesion area meets the second threshold condition, it is marked as a confirmed lesion area, otherwise, it is marked as a suspected lesion area;

[0096] S403: The third classifier outputs a second detection result including a suspected lesion area and a confirmed lesion area.

[0097] Preferably, Figure 6 As shown, the image to be sorted is distinguished and marked according to the normal area, the suspected lesion area and the confirmed lesion area, and the marked image to be sorted is visually displayed, including:

[0098] S501: Color marking the normal areas, suspected lesion areas and confirmed lesion areas classified in the image to be sorted;

[0099] S502: Performing a visualization operation on the color-labeled image to be sorted. Embodiment 2

[0100] The present invention also provides an image-assisted sorting system for gastroenterology, which can at least be used to implement the sorting method described in the first embodiment. Figure 7 As shown, including:

[0101] An image processing module 100, for acquiring and processing the image to be sorted corresponding to the digestive tract;

[0102] A first classification module 200, for classifying the image to be sorted according to the digestive structure using a first classifier, and outputting sub-images corresponding to different digestive structures, wherein the first classifier is trained based on sample images;

[0103] The second classification module 300 is used to perform lesion detection on all sub-images using a second classifier based on preset condition 1 to obtain normal areas and lesion areas;

[0104] The third classification module 400 is used to perform secondary lesion detection on the lesion area using a third classifier based on the preset condition 2 to obtain a suspected lesion area and a confirmed lesion area;

[0105] The identification module 500 is used to distinguish and identify the image to be sorted according to the normal area, the suspected lesion area and the confirmed lesion area, and to visually display the identified image to be sorted.

[0106] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0107] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically-erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0108] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A gastroenterology image-assisted arrangement method, characterized in that: The finishing method comprises the following steps: Acquiring and processing the image to be sorted corresponding to the digestive tract; Using a first classifier to classify the image to be sorted according to the digestive structure, and outputting sub-images corresponding to different digestive structures, wherein the first classifier is trained based on sample images; Based on the preset condition 1, the second classifier is used to perform lesion detection on all sub-images to obtain normal areas and lesion areas; Based on the second preset condition, a third classifier is used to perform secondary lesion detection on the lesion area to obtain a suspected lesion area and a confirmed lesion area; Differentiating and marking the image to be sorted according to the normal area, the suspected lesion area and the confirmed lesion area, and visually displaying the marked image to be sorted; The method of using the first classifier to classify the image to be sorted according to the digestive structure and outputting sub-images including different digestive structures includes: Segmenting the image to be sorted based on an edge detection method to obtain a segmented image including different digestive structures; Extracting features of each segmented image based on a Gabor filter, wherein the features include shape, size, texture, and boundary curvature; Based on the features of the segmented image, the segmented image is identified and classified using a trained first classifier, and sub-images corresponding to different digestive structures are output.

2. The image-assisted arrangement method for gastroenterology according to claim 1, characterized in that: The image to be sorted and the sample image are one of an endoscopic image, an X-ray angiography image, a CT image and an MRI image of the digestive tract.

3. The image-assisted arrangement method for gastroenterology according to claim 2, characterized in that: The second classifier and the third classifier each include sub-classification modules corresponding to different digestive structures, wherein the digestive structures include stomach, liver, and small intestine.

4. The image-assisted arrangement method for gastroenterology according to claim 1, characterized in that: The step of acquiring and processing the image to be sorted corresponding to the digestive tract includes: Acquire an image to be sorted corresponding to the digestive tract; The image to be sorted is preprocessed, wherein the preprocessing includes noise reduction, artifact removal, and contrast enhancement.

5. The image-assisted arrangement method for gastroenterology according to claim 4, characterized in that: The training process of the first classifier includes: Collecting sample images, wherein each sample image includes a sample digestion structure and a sample feature corresponding to the sample digestion structure; Using the sample features of each sample image as input and the sample digestion structure as output to train a first classifier, wherein the first classifier adopts a deep learning framework and is built based on a neural network; Output the digestive structure label corresponding to each sample image.

6. The image-assisted arrangement method for gastroenterology according to claim 1, characterized in that: Based on the preset condition 1, the second classifier is used to perform lesion detection on all sub-images to obtain normal areas and lesion areas, including: Setting a preset condition 1, wherein the preset condition 1 is a first threshold condition; All sub-images are input into the second classifier for detection in sequence. If the sub-image meets the first threshold condition, it is marked as a diseased area, otherwise, it is marked as a normal area; The second classifier outputs a first detection result including a normal area and a lesion area.

7. The image-assisted arrangement method for gastroenterology according to claim 1, characterized in that: Based on the second preset condition, the third classifier is used to perform secondary lesion detection on the lesion area to obtain the suspected lesion area and the confirmed lesion area, including: Setting a second preset condition, wherein the second preset condition is a second threshold condition; All lesion areas are input into the third classifier for detection in sequence. If the lesion area meets the second threshold condition, it is marked as a confirmed lesion area, otherwise, it is marked as a suspected lesion area; The third classifier outputs a second detection result including a suspected lesion area and a confirmed lesion area.

8. The image-assisted arrangement method for gastroenterology according to claim 1, characterized in that: The step of distinguishing and marking the image to be sorted according to the normal area, the suspected lesion area and the confirmed lesion area, and visually displaying the marked image to be sorted includes: Color marking is performed on the normal areas, suspected lesion areas and confirmed lesion areas classified in the images to be sorted; Perform visualization operations on the images to be sorted after color marking.

9. An image-assisted arrangement system for gastroenterology, characterized in that: include: An image processing module, used for acquiring and processing the image to be sorted corresponding to the digestive tract; A first classification module, used to classify the image to be sorted according to the digestive structure using a first classifier, and output sub-images corresponding to different digestive structures, wherein the first classifier is trained based on sample images; The second classification module is used to perform lesion detection on all sub-images based on preset condition 1 using the second classifier to obtain normal areas and lesion areas; A third classification module is used to perform secondary lesion detection on the lesion area using a third classifier based on preset condition 2 to obtain a suspected lesion area and a confirmed lesion area; An identification module, used to distinguish and identify the image to be sorted according to the normal area, the suspected lesion area and the confirmed lesion area, and to visually display the identified image to be sorted; The method of using the first classifier to classify the image to be sorted according to the digestive structure and outputting sub-images including different digestive structures includes: Segmenting the image to be sorted based on an edge detection method to obtain a segmented image including different digestive structures; Extracting features of each segmented image based on a Gabor filter, wherein the features include shape, size, texture, and boundary curvature; Based on the features of the segmented image, the segmented image is identified and classified using a trained first classifier, and sub-images corresponding to different digestive structures are output.

Images