Evaluation Method, Device, Electronic Device and Storage Medium for Image Segmentation Quality

Abstract

The present disclosure provides a method, an apparatus, an electronic device, and a computer-readable storage medium for evaluating the quality of image segmentation, which relates to the field of artificial intelligence. Among them, the method for evaluating the quality of image segmentation includes: segmenting an image based on a level set segmentation model to obtain a segmentation result; determining a metric parameter of pixels in the image based on the level set signed distance function; determining a warning value of the pixels in the image based on the metric parameter; obtaining a warning area of the image according to the warning value; and generating a quality evaluation result of the segmentation result based on the area of the warning area. Through the technical solution of the present disclosure, a relatively accurate detection result can be obtained, and the solution has good robustness. Furthermore, it is beneficial to further screen out reliable areas and warning areas in the segmentation result, thereby providing a reliable reference for the overall segmentation quality of the image.

Description

Background Art

[0002] An independent ground print is a ground print with an independent geometric boundary structure outside the lane lines on the road, such as direction arrows, diamond deceleration signs, etc. The geometric boundary information of the independent ground print is an important part of the high-precision map, which can provide key reference information for the autonomous driving navigation algorithm.

[0003] In the related art, although deep learning algorithms can be used for the segmentation and extraction of independent ground prints, the accuracy and robustness of the segmentation results cannot be accurately evaluated.

[0004] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of the present disclosure, and thus may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0005] The purpose of the present disclosure is to provide a method for evaluating the quality of image segmentation, a device for evaluating the quality of image segmentation, an electronic device, and a computer-readable storage medium, which can at least to some extent realize a reliable evaluation of the quality of segmentation results such as independent ground prints segmented from an image based on a segmentation model.

[0006] Other features and advantages of the present disclosure will become apparent through the following detailed description, or will be learned in part through the practice of the present disclosure.

[0007] According to one aspect of the present disclosure, there is provided a method for evaluating the quality of image segmentation, including: segmenting an image based on a level set segmentation model to obtain a segmentation result; determining a metric parameter of a pixel in the image based on a level set signed distance function; determining a warning value of the pixel in the image based on the metric parameter; obtaining a warning area of the image according to the warning value; and generating a quality evaluation result of the segmentation result based on the area of the warning area.

[0008] In one embodiment, the obtaining the warning area of the image according to the warning value includes: determining a first threshold according to the warning value; and dividing the warning area in the image according to the first threshold.

[0009] In one embodiment, the determining the first threshold according to the warning value includes: calculating the median of the warning values of a plurality of the pixels; and using the median as the first threshold.

[0010] In one embodiment, the dividing the warning area in the image according to the first threshold includes: dividing the area where the warning value is greater than the first threshold into the warning area.

[0011] In one embodiment, obtaining the warning value of the image pixel based on the metric parameter includes: obtaining the warning value of the pixel according to the reciprocal of the absolute value of the metric parameter of the pixel.

[0012] In one embodiment, generating the evaluation result of the image segmentation quality based on the area of the warning region includes: calculating the area of the warning region and taking it as the first area; calculating the area of the segmentation result in the image and taking it as the second area; generating the evaluation result based on the relative size relationship or proportional relationship between the first area and the second area.

[0013] In one embodiment, calculating the area of the warning region and taking it as the first area includes: determining the number of pixels whose warning value is greater than the first threshold; determining the first area according to the number of pixels.

[0014] In one embodiment, calculating the area of the segmentation result in the image and taking it as the second area includes: forming a circumscribed rectangle based on the segmentation result of the image; determining the area of the circumscribed rectangle as the second area.

[0015] In one embodiment, generating the evaluation result based on the relative size relationship or proportional relationship between the first area and the second area includes: determining whether the segmentation result of the image is available or unavailable according to the relationship between the ratio of the first area to the second area and the second threshold.

[0016] According to another aspect of the present disclosure, there is provided an apparatus for evaluating the quality of image segmentation, including: a segmentation module for segmenting an image based on a level set segmentation model to obtain a segmentation result; a first determination module for determining a metric parameter of a pixel in the image based on a level set signed distance function; a second determination module for determining a warning value of a pixel in the image based on the metric parameter; an acquisition module for obtaining a warning region of the image according to the warning value; a generation module for generating a quality evaluation result of the segmentation result based on the area of the warning region.

[0017] According to still another aspect of the present disclosure, there is provided an electronic device, including: a processor; and a memory for storing executable instructions of the processor; wherein, the processor is configured to execute the method for evaluating the quality of image segmentation according to any one of the above via executing the executable instructions.

[0018] According to yet another aspect of the present disclosure, there is provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for evaluating the quality of image segmentation according to any one of the above is implemented.

[0019] The evaluation scheme for the quality of image segmentation provided by the embodiments of the present disclosure uses a level set segmentation model to segment an image in a manner based on the evolution of the level set curve. When the curve evolution stops to obtain the segmentation result, the signed distance function representing the level set is used to calculate the distance between the pixel points of the image and the boundary curve of the segmentation result to obtain a metric parameter, so as to represent the distance between the pixel points and the boundary curve based on the metric parameter. Further, an early warning value for measuring the segmentation quality is determined based on the metric parameter. After generating an early warning area according to the early warning value, the evaluation result of the segmentation quality is obtained according to the area of the early warning area. Based on this scheme, the quality evaluation of the segmentation result is realized. This quality evaluation scheme has good robustness and can obtain relatively accurate evaluation results, which is conducive to further screening out reliable areas and early warning areas in the segmentation result, thereby providing a reliable reference for the overall segmentation quality of the image.

[0020] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure. Obviously, the accompanying drawings in the following description are only some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0022] Figure 1 A schematic diagram showing the structure of an evaluation system for the quality of image segmentation in an embodiment of the present disclosure;

[0023] Figure 2 A flowchart showing a method for evaluating the quality of image segmentation in an embodiment of the present disclosure;

[0024] Figure 3 A flowchart showing a method for obtaining an early warning area in an embodiment of the present disclosure;

[0025] Figure 4 A flowchart showing another method for evaluating the quality of image segmentation in an embodiment of the present disclosure;

[0026] Figure 5 A flowchart showing still another method for evaluating the quality of image segmentation in an embodiment of the present disclosure;

[0027] Figure 6 A schematic diagram showing the segmentation result of the level set segmentation model in an embodiment of the present disclosure;

[0028] Figure 7Schematic diagram showing the segmentation error area in the segmentation result of the level set segmentation model in an embodiment of the present disclosure;

[0029] Figure 8 Schematic diagram showing the warning area obtained in an embodiment of the present disclosure;

[0030] Figure 9 Schematic diagram showing an evaluation device for the quality of an image segmentation in an embodiment of the present disclosure;

[0031] Figure 10 Schematic diagram showing an electronic device in an embodiment of the present disclosure. Detailed implementation manners

[0032] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be more complete and comprehensive, and will fully convey the concept of the example embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0033] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0034] The solution provided by the present application determines a warning value for measuring the segmentation quality based on a metric parameter, so that after generating a warning area according to the warning value, an evaluation result of the segmentation quality can be obtained based on the area of the warning area. The quality evaluation of the segmentation result is realized based on this solution. This quality evaluation solution has good robustness and can obtain relatively accurate evaluation results, which is beneficial to further screening out reliable areas and warning areas in the segmentation result, thereby providing a reliable reference for the overall segmentation quality of the image.

[0035] For ease of understanding, several terms related to the present application will be explained first below.

[0036] Artificial Intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a way similar to human intelligence. Research in this field includes robots, speech recognition, image recognition, natural language processing, and expert systems, etc. Since its inception, the theory and technology of artificial intelligence have become increasingly mature, and the application fields have also been continuously expanding. It can be envisioned that future technological products brought about by artificial intelligence will be "containers" of human wisdom. Artificial intelligence can simulate the information processes of human consciousness and thinking. Artificial intelligence is not human intelligence, but it can think like humans and may even exceed human intelligence.

[0037] As a method to achieve artificial intelligence, machine learning is mainly used to analyze data, learn from it, and then make decisions and predictions about events in the real world, including decision tree learning, inductive logic programming, clustering, classification, regression, reinforcement learning, and Bayesian networks, etc. Deep learning, as a technology to achieve machine learning, includes artificial neural network model algorithms, etc.

[0038] An independent ground print is a ground print in the road that has an independent geometric boundary structure different from the lane lines, such as direction arrows, diamond deceleration signs, etc.

[0039] The solution provided in the embodiments of this application involves technologies such as neural network modeling and machine learning, and will be specifically described through the following embodiments.

[0040] Figure 1 The structural schematic diagram of a parking lot traffic system in the embodiments of the present disclosure is shown, including a plurality of terminals 120 and a server cluster 140.

[0041] The terminal 120 can be a mobile terminal such as a mobile phone, game console, tablet computer, e-book reader, smart glasses, MP4 (Moving Picture Experts Group Audio Layer IV) player, smart home device, AR (Augmented Reality) device, VR (Virtual Reality) device, etc. Alternatively, the terminal 120 can also be a personal computer (PC), such as a laptop computer and a desktop computer, etc.

[0042] Among them, an application program for providing parking lot traffic can be installed in the terminal 120.

[0043] The terminal 120 is connected to the server cluster 140 through a communication network. Optionally, the communication network is a wired network or a wireless network.

[0044] The server cluster 140 is a single server, or consists of several servers, or is a virtualization platform, or is a cloud computing service center. The server cluster 140 is used to provide background services for the evaluation application for image segmentation quality and the training application for traffic prediction models. Optionally, the server cluster 140 undertakes the main computing work, and the terminal 120 undertakes the secondary computing work; or, the server cluster 140 undertakes the secondary computing work, and the terminal 120 undertakes the main computing work; or, a distributed computing architecture is adopted between the terminal 120 and the server cluster 140 for collaborative computing.

[0045] In some alternative embodiments, the server cluster 140 is used to store the evaluation model for image segmentation quality, the prediction method, etc.

[0046] Optionally, the clients of the applications installed in different terminals 120 are the same, or the clients of the applications installed on two terminals 120 are clients of the same type of application on different control system platforms. Based on the differences in terminal platforms, the specific form of the application client can also be different. For example, the application client can be a mobile phone client, a PC client, or a World Wide Web (Web) client, etc.

[0047] Those skilled in the art can know that the number of the above-mentioned terminals 120 can be more or less. For example, the above-mentioned terminal can be only one, or there can be dozens or hundreds of the above-mentioned terminals, or even more. The embodiments of the present application do not limit the number and device type of the terminals.

[0048] Optionally, the system may further include a management device ( Figure 1 (not shown), which is connected to the server cluster 140 through a communication network. Optionally, the communication network is a wired network or a wireless network.

[0049] Optionally, the above-mentioned wireless network or wired network uses standard communication technologies and / or protocols. The network is usually the Internet, but can also be any network, including but not limited to any combination of a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a private network or a virtual private network). In some embodiments, technologies and / or formats including Hyper Text Mark-up Language (HTML), Extensible Markup Language (XML), etc. are used to represent the data exchanged through the network. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec), etc. can be used to encrypt all or some of the links. In other embodiments, customized and / or dedicated data communication technologies can be used to replace or supplement the above data communication technologies.

[0050] Next, each step in the method for evaluating the image segmentation quality and the method for training the traffic prediction model in the present exemplary embodiment will be described in more detail with reference to the accompanying drawings and embodiments.

[0051] Figure 2 The flowchart of a method for evaluating the image segmentation quality in an embodiment of the present disclosure is shown. The method provided by the embodiment of the present disclosure can be executed by any electronic device with computing and processing capabilities, such as the Figure 1 terminal 120 and / or the server cluster 140 in. In the following illustrative examples, the terminal 120 is used as the execution entity for illustrative purposes.

[0052] As Figure 2 shown, the terminal 120 executes the method for evaluating the image segmentation quality, including the following steps:

[0053] Step S202, segment the image based on the level set segmentation model to obtain a segmentation result.

[0054] Among them, the level set segmentation model adopts the level set segmentation method. The level set is a digital method for tracking the movement of contours and surfaces. By setting the contour as the zero level set of a high-dimensional function, the high-dimensional function is the level set function. By differentiating the level set function, the zero level set is extracted from the output to obtain the moving contour, that is, the segmentation result to be segmented in the image.

[0055] Step S204, determining the metric parameter of the pixels in the image based on the level set signed distance function;

[0056] Among them, the metric parameter is the shortest distance from the point to the boundary curve of the segmentation result represented by the signed distance function (SDF). The sign of the distance depends on whether the point is inside or outside the boundary curve. Usually, a positive sign is taken outside and a negative sign is taken inside.

[0057] Specifically, the numerical method based on the level set represents the movement and change of the surface in an implicit way. All information related to the surface movement and the situation of the moving surface itself are reflected in the level set function. This method reduces the calculation area from the entire image plane to a narrow band around the curve, greatly reducing the calculation amount and improving the efficiency.

[0058] Step S206, determining the warning value of the pixels in the image based on the metric parameter.

[0059] Among them, the warning value can be determined based on the quality of the image and the severity of the evaluation of the segmentation quality.

[0060] Step S208, obtaining the warning area of the image according to the warning value.

[0061] Among them, the warning area refers to evaluating the quality of the segmentation result of the image and providing information about the areas where the deep learning algorithm may fail in advance in the high-precision map vectorization link and the manual quality inspection link of the image segmentation result.

[0062] Step S210, generating a quality evaluation result of the segmentation result based on the area of the warning area.

[0063] Among them, those skilled in the art can understand that the larger the area of the warning area, the worse the quality of the segmentation result, and the smaller the area of the warning area, the better the quality of the segmentation result.

[0064] In this embodiment, by adopting a level set segmentation model, the image is segmented in a manner based on the evolution of the level set curve. While the curve evolution stops to obtain the segmentation result, the signed distance function representing the level set is used to calculate the distance between the pixel points of the image and the boundary curve of the segmentation result to obtain a metric parameter, so as to represent the distance between the pixel points and the boundary curve based on the metric parameter. Further, an early warning value for measuring the segmentation quality is determined based on the metric parameter. After generating an early warning area according to the early warning value, the evaluation result of the segmentation quality is obtained according to the area of the early warning area. The quality evaluation of the segmentation result is realized based on this solution. This quality evaluation scheme has good robustness and can obtain relatively accurate evaluation results, which is conducive to further screening out reliable areas and early warning areas in the segmentation result, thereby providing a reliable reference for the overall segmentation quality of the image.

[0065] In one embodiment, step S206 of obtaining the early warning value of the image pixels based on the metric parameter includes:

[0066] Obtaining the early warning value of the pixel according to the reciprocal of the absolute value of the metric parameter of the pixel.

[0067] In this embodiment, the value obtained by the signed distance function is used as the metric parameter, and the absolute value calculation is performed based on the pixel points. After normalizing the obtained absolute value, the reciprocal operation is performed on it, and the reciprocal value is used as the early warning value, so as to establish an association relationship between the level set signed distance function, that is, the metric parameter, and the early warning value, thereby ensuring the possibility of the quality evaluation of the segmentation result. Among them, the larger the early warning value, the farther the pixel point is from the segmentation boundary, indicating that the segmentation result is less reliable, and the smaller it is, the more reliable the segmentation result is.

[0068] In one embodiment, step S208 of obtaining the early warning area of the image according to the early warning value includes: determining a first threshold according to the early warning value.

[0069] Among them, the first threshold is a threshold for measuring whether the pixel points in the image belong to the early warning area.

[0070] As Figure 3 shown, in one embodiment, as an alternative implementation, determining the first threshold according to the early warning value includes:

[0071] Step S302, calculating the median of the early warning values of multiple pixels.

[0072] Among them, the median of the early warning values of multiple pixels can be understood as the average value of the early warning values of multiple pixels.

[0073] Step S304, taking the median as the first threshold.

[0074] In this embodiment, the median value is determined as the first threshold. On the one hand, it can ensure the rationality of detecting the warning area. On the other hand, dividing the warning area based on the first threshold can also ensure the reliability of the obtained warning area.

[0075] Furthermore, the first threshold can be dynamically adjusted based on the source of the image and the evaluation criteria for the segmentation quality. Specifically, initial parameters are used to measure the source of the image, and evaluation parameters are used to evaluate the segmentation quality. For example, compared with the image captured by a camera and the image intercepted from a video, the former has higher image quality. Therefore, correspondingly, the evaluation criteria for its segmentation result are higher.

[0076] After determining the first threshold according to the warning value, further, it also includes: dividing the warning area in the image according to the first threshold.

[0077] In one embodiment, as an optional implementation manner, dividing the warning area in the image according to the first threshold includes:

[0078] Step S306, dividing the area where the warning value is greater than the first threshold into the warning area.

[0079] In this embodiment, the metric parameter is obtained according to the signed distance function, and the warning value of each pixel in the segmentation area is calculated based on the absolute value of the metric parameter. After determining the first threshold, based on this evaluation method, for the detected area where the warning value is greater than the first threshold, it can be understood as the area with poor segmentation quality. The detected area where the warning value is greater than the first threshold is divided into the warning area, so that the evaluation of the segmentation result can be carried out based on the obtained warning area to ensure the feasibility and reliability of the segmentation result evaluation.

[0080] As Figure 4 shown, in one embodiment, step S210 generating the evaluation result of the image segmentation quality based on the area of the warning area includes:

[0081] Step S402, calculating the area of the warning area and taking it as the first area.

[0082] Among them, taking the area of the warning area as the detection object can obtain the detection object without considering factors such as the segmentation shape.

[0083] Step S404, calculating the area of the segmentation result in the image and taking it as the second area.

[0084] Among them, taking the area of the segmentation result as the detection benchmark is a direct evaluation of the segmentation result and does not require taking the determined segmentation contour as a reference. Therefore, the detection method can be more general.

[0085] Step S406: Generate an evaluation result based on the relative magnitude relationship or proportional relationship between the first area and the second area.

[0086] In this embodiment, by detecting the relationship between the first area and the second area and determining the detection result as the evaluation result, a detection method for determining the segmentation quality of the evaluation result based on the relationship between the area of the warning region and the area of the segmentation result is realized. This method can obtain the evaluation result without obtaining the specific segmentation contour and the specific shape of the warning region, and can further output an image of the available segmentation result based on the evaluation result, and mark the unavailable image part and the region of failed segmentation to realize the effective utilization of the segmentation result.

[0087] Specifically, the evaluation result can be obtained based on the relative magnitude relationship between the first area and the second area. Taking the second area as the benchmark, the larger the first area, the worse the segmentation quality, and the smaller the first area, the better the segmentation quality.

[0088] In addition, the evaluation result can also be obtained based on the proportional relationship between the first area and the second area. The larger the proportional value, the worse the segmentation quality, and the smaller the proportional value, the better the segmentation quality.

[0089] In one embodiment, the specific implementation manner of step S402 for calculating the area of the warning region and using it as the first area includes:

[0090] Determine the number of pixels whose warning value is greater than the first threshold.

[0091] Determine the first area according to the number of pixels.

[0092] In this embodiment, as a specific calculation method for the area of the warning region, based on the relationship between the warning value and the first threshold, determine the number of pixels whose warning value is greater than the first threshold, so as to obtain the area of the warning region, that is, the first area, based on the number of pixels, in order to obtain the detection object for detecting the segmentation quality.

[0093] In one embodiment, the specific implementation manner of step S404 for calculating the area of the segmentation result in the image and using it as the second area includes:

[0094] Form a circumscribed rectangle based on the segmentation result of the image.

[0095] Determine the area of the circumscribed rectangle as the second area.

[0096] In this embodiment, through the four extreme points of the area to be segmented by the auxiliary segmentation tool, including the first extreme point A, the second extreme point B, the third extreme point C, and the fourth extreme point D, the coordinate information of these four extreme points is specifically carried in the object annotation instruction. Thus, according to the object annotation instruction, a segmented image corresponding to the image to be processed, that is, the segmentation result, is generated. The segmentation result includes the area formed by the first extreme point A, the second extreme point B, the third extreme point C, and the fourth extreme point D, that is, the circumscribed rectangle. Through the coordinate information of the four extreme points, the area of the circumscribed rectangle, that is, the second area, is obtained. Taking the second area as the comparison benchmark, the first area is compared with the second area to obtain the evaluation result, thereby realizing the reliable evaluation of the segmentation result.

[0097] In one embodiment, the specific implementation manner of generating the evaluation result based on the relative size relationship or proportional relationship between the first area and the second area in step S406 includes: determining whether the segmentation result of the image is available or unavailable according to the relationship between the ratio of the first area to the second area and the second threshold.

[0098] Among them, the second threshold is used as the benchmark for whether the segmentation result of the image is available.

[0099] In this embodiment, as a preferred calculation method, by setting the second threshold, the evaluation result is determined according to the relationship between the ratio of the first area to the second area and the second threshold. Specifically, if the ratio is greater than the second threshold, it indicates that the segmentation quality is poor; if the ratio is less than or equal to the second threshold, it indicates that the segmentation quality is good. Thus, the quantitative evaluation of the segmentation result is realized, and the evaluation method can ensure good robustness and generality.

[0100] In addition, those skilled in the art can understand that the second threshold can also be dynamically adjusted based on parameters such as the source of the image and the evaluation standard for the segmentation quality.

[0101] Such as Figure 5 , according to the evaluation method for the image segmentation quality of another embodiment of the present disclosure, it includes:

[0102] Step S502, obtaining the image to be segmented.

[0103] For example, taking the raster image as the image to be segmented, the raster image can be a road image collected by a camera or a laser point, and is generated through a vision algorithm. The raster image has characteristics such as high precision and rich ground object elements, and is very suitable for the production of high-precision maps.

[0104] Step S504, based on the image segmentation model of the level set, performing a segmentation operation on the image and obtaining the segmentation result.

[0105] Among them, as an algorithm in the field of active contour models, the image segmentation model has been widely verified in medical images in terms of its segmentation and detection effects.

[0106] Step S506, when the evolution of the level set curve in the image region stops, based on the segmentation result, obtain the value of the signed distance function representing the level set of the pixels in the image at this time as a metric parameter.

[0107] Step S508, perform an absolute value calculation on the metric parameter.

[0108] Step S510, perform a normalization operation on the absolute value, and then use its reciprocal value as the warning value.

[0109] Among them, the larger the warning value, the less reliable the segmentation result, and the smaller the warning value, the more reliable the segmentation result.

[0110] Step S512, obtain the first threshold according to the warning value.

[0111] Among them, the acquisition of the first threshold is mainly determined by the actual image source and the severity of the evaluation of the segmentation quality, and it is dynamically changing.

[0112] Step S514, divide the warning area based on the warning value and the first threshold.

[0113] Step S516, according to the warning area, calculate the relative area size of the area of the warning area and the area of the segmentation result to determine whether the overall result of the segmentation target is available.

[0114] Filter out images with different segmentation qualities according to whether the whole of the target to be segmented is available. For the vectorization operation of the segmentation result, more flexible vectorization parameter selection can be adopted with reference to the relative size of the warning area and the warning value of each pixel, so as to improve the vectorization accuracy.

[0115] In this implementation, the technical solution of the present disclosure divides the image segmentation warning algorithm into two stages, namely the level set segmentation model and the calculation of the warning area.

[0116] Specifically, first use the level set segmentation model to segment the image, then calculate the absolute value of the signed distance function representing the level set when the level set evolution curve stops, perform a normalization operation on the absolute value, and then use its reciprocal value as its warning value. This stage mainly establishes the connection between the segmentation result and the warning value.

[0117] Among them, Figure 6 shows the segmentation result obtained by segmenting the image using the level set segmentation model.

[0118] Further, after the segmentation is completed based on the level set segmentation model, a distance function representing the sign of the level set is obtained. The warning value of each pixel in the segmented region is calculated according to the absolute value of the sign distance function, the division threshold is obtained according to the size of the warning value, the area of the warning region is calculated according to the threshold, and then whether the overall result of the segmented object is available is determined according to the size relationship between the area of the warning region and the relative area of the region to be segmented. Finally, the available segmentation result image is output, and the unavailable images and the regions where the segmentation fails are marked.

[0119] Among them, Figure 7 shows the region where the segmentation error occurs, Figure 8 shows the warning region obtained based on the solution of the present disclosure. By Figure 7 and Figure 8 the comparison results, it can be determined that the evaluation solution of the present disclosure has high reliability and accuracy.

[0120] After going through the processes of the above two modules, the warning algorithm for the image segmentation result has been completely completed. This warning algorithm can achieve the warning of the segmentation failure region, provide reference information for the vectorization link in the high-precision map production, and at the same time can automatically perform quality inspection on whether the image quality meets the segmentation requirements of the algorithm, which will greatly reduce the subsequent manual quality inspection time of the images. When the number of images to be quality inspected is less than a certain proportion (according to the segmentation accuracy requirements of the customer delivery project), this part of the manual quality inspection link can even be removed.

[0121] The above method has shown good algorithm robustness and the accuracy of the warning region results through the warning experiment on the raster image of the independent ground printed matter, and can automatically generate the screening of the warning region and the image segmentation quality.

[0122] It should be noted that the above drawings are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present invention, rather than for limiting purposes. It is easy to understand that the processes shown in the above drawings do not indicate or limit the time sequence of these processes. Additionally, it is also easy to understand that these processes can be executed synchronously or asynchronously in, for example, multiple modules.

[0123] Those skilled in the art to which the present invention pertains can understand that various aspects of the present invention can be implemented as a system, method, or program product. Therefore, various aspects of the present invention can be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which can be collectively referred to as "circuit", "module", or "system" here.

[0124] Next, refer to Figure 9 to describe the evaluation device 900 for the image segmentation quality according to this embodiment of the present invention. Figure 9The evaluation device 400 for the image segmentation quality shown is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of the present invention.

[0125] The evaluation device 900 for the image segmentation quality is presented in the form of a hardware module. The components of the evaluation device 900 for the image segmentation quality may include but are not limited to: a segmentation module 902 for segmenting an image based on a level set segmentation model to obtain a segmentation result; a first determination module 904 for determining the metric parameter of the pixels in the image based on the level set signed distance function; a second determination module 906 for determining the warning value of the pixels in the image based on the metric parameter; an acquisition module 908 for obtaining the warning area of the image according to the warning value; and a generation module 910 for generating a quality evaluation result of the segmentation result based on the area of the warning area.

[0126] In one embodiment, the acquisition module 908 is further configured to: determine a first threshold according to the warning value; and divide the warning area in the image according to the first threshold.

[0127] In one embodiment, the acquisition module 908 is further configured to: calculate the median of the warning values of multiple pixels; and use the median as the first threshold.

[0128] In one embodiment, the acquisition module 908 is further configured to: divide the area where the warning value is greater than the first threshold into the warning area.

[0129] In one embodiment, the second determination module 906 is further configured to: obtain the warning value of the pixel according to the reciprocal of the absolute value of the metric parameter of the pixel.

[0130] In one embodiment, the generation module 910 is further configured to: calculate the area of the warning area as the first area; calculate the area of the segmentation result in the image as the second area; and generate an evaluation result based on the relative size relationship or proportional relationship between the first area and the second area.

[0131] In one embodiment, the generation module 910 is further configured to: determine the number of pixels whose warning value is greater than the first threshold; and determine the first area according to the number of pixels.

[0132] In one embodiment, the generation module 910 is further configured to: form a circumscribed rectangle based on the segmentation result of the image; and determine the area of the circumscribed rectangle as the second area.

[0133] In one embodiment, the generation module 910 is further configured to: determine whether the segmentation result of the image is available or unavailable according to the relationship between the ratio of the first area to the second area and the second threshold.

[0134] Next, refer to Figure 10 to describe the electronic device 1000 according to this embodiment of the present invention. Figure 10The illustrated electronic device 1000 is merely an example and shall not impose any limitation on the functions and scope of use of the embodiments of the present invention.

[0135] As Figure 10 shown, the electronic device 1000 is presented in the form of a general-purpose computing device. The components of the electronic device 1000 may include, but are not limited to: at least one of the above-mentioned processing units 1010, at least one of the above-mentioned storage units 1020, and a bus 1030 connecting different system components (including the storage unit 1020 and the processing unit 1010).

[0136] Among them, the storage unit stores program code, which can be executed by the processing unit 1010, so that the processing unit 1010 executes the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of this specification. For example, the processing unit 1010 can execute steps S202, S204, S206, and S208 as Figure 2 shown, as well as other steps defined in the evaluation method of the image segmentation quality and / or the training method of the traffic prediction model disclosed in the present disclosure.

[0137] The storage unit 1020 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 10201 and / or a cache storage unit 10202, and may further include a read-only storage unit (ROM) 10203.

[0138] The storage unit 1020 may further include a program / utilities 10204 having a set (at least one) of program modules 10205. Such program modules 10205 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. The implementation of a network environment may be included in each or some combination of these examples.

[0139] The bus 1030 may represent one or more of several types of bus structures, including a storage unit bus or a storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any bus structure in a variety of bus structures.

[0140] The electronic device 1000 can also communicate with one or more external devices 1060 (such as a keyboard, a pointing device, a Bluetooth device, etc.), and can also communicate with one or more devices that enable a user to interact with the electronic device, and / or communicate with any device (such as a router, a modem, etc.) that enables the electronic device 1000 to communicate with one or more other computing devices. Such communication can be carried out through the input / output (I / O) interface 1050. Moreover, the electronic device 1000 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through the network adapter 1050. As shown in the figure, the network adapter 1050 communicates with other modules of the electronic device 1000 through the bus 1030. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in combination with the electronic device, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0141] Through the description of the above embodiments, those skilled in the art can easily understand that the exemplary embodiments described herein can be implemented by software, or can be implemented by the way of software combined with necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product, and the software product can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

[0142] In an exemplary embodiment of the present disclosure, there is also provided a computer-readable storage medium, on which a program product capable of implementing the above method of this specification is stored. In some possible implementation manners, various aspects of the present invention can also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of this specification.

[0143] The program product for implementing the above method according to the embodiments of the present invention can adopt a portable compact disc read-only memory (CD-ROM) and include program code, and can run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, the readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, apparatus, or device.

[0144] A computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, carrying readable program code. Such a propagated data signal may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. The readable signal medium may also be any readable medium other than a readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device.

[0145] The program code contained on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical fiber cable, RF, and the like, or any suitable combination of the foregoing.

[0146] The program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and also including conventional procedural programming languages such as the "C" language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0147] It should be noted that although several modules or units of a device for action execution are mentioned in the foregoing detailed description, such a division is not mandatory. In fact, according to an embodiment of the present disclosure, the features and functions of two or more of the above-described modules or units may be embodied in one module or unit. Conversely, the features and functions of one module or unit described above may be further divided and embodied by a plurality of modules or units.

[0148] In addition, although the steps of the methods in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in that specific order, or that all of the steps shown must be performed to achieve the desired result. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step for execution, and / or one step may be decomposed into multiple steps for execution, etc.

[0149] From the description of the above embodiments, those skilled in the art can easily understand that the example embodiments described herein can be implemented by software or by a combination of software and necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (such as a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.

[0150] After considering the specification and practicing the invention disclosed herein, those skilled in the art will readily conceive of other embodiments of the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and examples are only regarded as exemplary, and the true scope and spirit of the present disclosure are pointed out by the appended claims.

Claims

1. A method for evaluating the quality of image segmentation, characterized in that Including: Segmenting an image based on a level set segmentation model to obtain a segmentation result; Determining a metric parameter of pixels in the image based on a level set signed distance function, where the metric parameter is obtained by calculating the distance between the pixels in the image and the boundary curve of the segmentation result; Obtaining a warning value for each pixel based on the reciprocal of the absolute value of the metric parameter of the pixel; Obtaining a warning area of the image according to the warning value; Generating a quality evaluation result of the segmentation result based on the area of the warning area.

2. The evaluation method for the quality of image segmentation according to claim 1, wherein, The obtaining the warning area of the image according to the warning value includes: Determining a first threshold according to the warning value; Dividing the warning area in the image according to the first threshold.

3. The evaluation method for the quality of image segmentation according to claim 2, wherein The determining the first threshold according to the warning value includes: Calculating the median of the warning values of multiple pixels; Taking the median as the first threshold.

4. The evaluation method for the quality of image segmentation according to claim 2, characterized in that, The dividing the warning area in the image according to the first threshold includes: Dividing the area where the warning value is greater than the first threshold into the warning area.

5. The evaluation method for the quality of image segmentation according to any one of claims 2 to 4, characterized in that, The generating the evaluation result of the image segmentation quality based on the area of the warning area includes: Calculating the area of the warning area and taking it as the first area; Calculating the area of the segmentation result in the image and taking it as the second area; Generating the evaluation result based on the relative size relationship or proportional relationship between the first area and the second area.

6. The evaluation method for the quality of image segmentation according to claim 5, characterized in that, The calculating the area of the warning area and taking it as the first area includes: Determining the number of pixels whose warning value is greater than the first threshold; Determining the first area according to the number of pixels.

7. The method for evaluating the quality of image segmentation according to claim 5, wherein The calculating the area of the segmentation result in the image and taking it as the second area includes: Forming a circumscribed rectangle based on the segmentation result of the image; Determining the area of the circumscribed rectangle as the second area.

8. The evaluation method for the quality of image segmentation according to claim 5, wherein The generating the evaluation result based on the relative size relationship or proportional relationship between the first area and the second area includes: Determining whether the segmentation result of the image is available or unavailable according to the relationship between the ratio of the first area to the second area and a second threshold.

9. An evaluation device for the quality of image segmentation, characterized in that, Including: A segmentation module for segmenting an image based on a level set segmentation model to obtain a segmentation result; A first determination module for determining a metric parameter of pixels in the image based on a level set signed distance function, where the metric parameter is obtained by calculating the distance between the pixels in the image and the boundary curve of the segmentation result; A second determination module for obtaining a warning value for each pixel based on the reciprocal of the absolute value of the metric parameter of the pixel; An acquisition module for obtaining the warning area of the image according to the warning value; A generation module for generating a quality evaluation result of the segmentation result based on the area of the warning area.

10. An electronic device, characterized in that, Including: A processor; And A memory for storing executable instructions of the processor; Wherein, the processor is configured to execute the method for evaluating the quality of image segmentation according to any one of claims 1 to 8 by executing the executable instructions.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for evaluating the quality of image segmentation according to any one of claims 1 to 8.

Images