CN114612367A - Evaluation method and device of image processing algorithm, computer equipment and storage medium - Google Patents
Evaluation method and device of image processing algorithm, computer equipment and storage medium Download PDFInfo
- Publication number
- CN114612367A CN114612367A CN202011401108.3A CN202011401108A CN114612367A CN 114612367 A CN114612367 A CN 114612367A CN 202011401108 A CN202011401108 A CN 202011401108A CN 114612367 A CN114612367 A CN 114612367A
- Authority
- CN
- China
- Prior art keywords
- image
- evaluation
- evaluation model
- image processing
- pixel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Library & Information Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The application relates to an evaluation method and device of an image processing algorithm, a computer device and a storage medium. The method comprises the following steps: acquiring reference data; processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein versions of the plurality of image processing algorithms are different; evaluating the plurality of processed images by utilizing an evaluation model to obtain an image evaluation result; wherein the image evaluation result is used to characterize differences between the plurality of processed images. By adopting the method, the maintenance difficulty can be reduced, and the iteration efficiency of the image processing algorithm can be improved.
Description
Technical Field
The present application relates to the field of image processing technologies, and in particular, to an evaluation method and apparatus for an image processing algorithm, a computer device, and a storage medium.
Background
The research and development of the image processing algorithm is an iterative process of continuous optimization and reconstruction. In the process, images processed by the image processing algorithms before and after iteration need to be frequently evaluated, so that the images are prevented from being degraded due to the iteration of the image processing algorithms. For example, ultrasound images processed by ultrasound image processing algorithms are evaluated.
In the related art, the evaluation mode is usually that human eyes observe image differences, but the evaluation mode depends on the experience of observers, and the evaluation result is not objective. In order to avoid the above problem, currently, an evaluation code is inserted into an algorithm code of an image processing algorithm, and the image processing algorithm is executed to obtain an image and an evaluation result at the same time.
However, in the above evaluation method, when the programmer iterates the image processing algorithm, the programmer needs to maintain the evaluation code and the algorithm code at the same time, which not only has high maintenance difficulty, but also reduces the iteration efficiency of the image processing algorithm.
Disclosure of Invention
In view of the foregoing, it is desirable to provide an image processing algorithm evaluation method, an image processing algorithm evaluation apparatus, a computer device, and a storage medium, which can reduce maintenance difficulty and improve iteration efficiency of an image processing algorithm.
A method of evaluating an image processing algorithm, the method comprising:
acquiring reference data;
respectively processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein the versions of the plurality of image processing algorithms are different;
evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result; wherein the image evaluation results are used to characterize differences between the plurality of processed images.
In one embodiment, the evaluating model includes a global image comparison evaluating model, and the evaluating the plurality of processed images by using the evaluating model to obtain the image evaluating result includes:
and inputting the plurality of processed images into a full-image comparison and evaluation model, and comparing the pixel value of each pixel point in the plurality of processed images by using the full-image comparison and evaluation model to obtain an image evaluation result output by the full-image comparison and evaluation model.
In one embodiment, the evaluating model includes a partition comparison evaluating model, and the evaluating the plurality of processed images by using the evaluating model to obtain the image evaluating result includes:
and inputting the plurality of processing images into the partition comparison and evaluation model, and comparing the pixel values of the corresponding partitions in the plurality of processing images by using the partition comparison and evaluation model to obtain an image evaluation result output by the partition comparison and evaluation model.
In one embodiment, the evaluating model includes a pixel interval evaluating model, the pixel interval evaluating model is provided with a plurality of pixel value intervals, and the evaluating the plurality of processed images by using the evaluating model to obtain the image evaluating result includes:
inputting the multiple processed images into a pixel interval evaluation model, and utilizing the pixel interval evaluation model to count the number of pixel points in each pixel value interval in the multiple processed images to obtain an image evaluation result output by the pixel interval evaluation model according to the statistical result.
In one embodiment, the evaluating model includes a custom evaluating model, at least one custom scheme is set in the custom evaluating model, and the evaluating of the plurality of processed images by using the evaluating model to obtain the image evaluating result includes:
and inputting the plurality of processing images into a user-defined evaluation model, and calculating by using the user-defined evaluation model according to the user-defined scheme and the plurality of processing images to obtain an image evaluation result output by the user-defined evaluation model.
In one embodiment, the evaluation model further comprises a performance evaluation model, and the method further comprises:
acquiring the running performance information of each image processing algorithm in the process of processing the reference data by a plurality of image processing algorithms;
and analyzing the operation performance information of each image processing algorithm by using the performance evaluation model to obtain a performance evaluation result.
In one embodiment, after analyzing the operation performance information of each image processing algorithm by using the performance evaluation model to obtain a performance evaluation result, the method further includes:
generating an evaluation report according to the image evaluation result and the performance evaluation result; the evaluation report includes an image evaluation result, a performance evaluation result, and a difference image generated from the plurality of processed images.
An apparatus for evaluating an image processing algorithm, the apparatus comprising:
the data acquisition module is used for acquiring reference data;
the image processing module is used for processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein the versions of the plurality of image processing algorithms are different;
the image evaluation module is used for evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result; wherein the image evaluation results are used to characterize differences between the plurality of processed images.
In one embodiment, the evaluation model includes a full-image comparison evaluation model, and the image evaluation module is specifically configured to input the multiple processed images into the full-image comparison evaluation model, and compare the pixel values of each pixel point in the multiple processed images by using the full-image comparison evaluation model to obtain an image evaluation result output by the full-image comparison evaluation model.
In one embodiment, the evaluation model includes a partition comparison evaluation model, and the image evaluation module is specifically configured to input the multiple processed images into the partition comparison evaluation model, and compare pixel values of corresponding partitions in the multiple processed images with the partition comparison evaluation model to obtain an image evaluation result output by the partition comparison evaluation model.
In one embodiment, the evaluation model includes a pixel interval evaluation model, and the image evaluation module is specifically configured to input the multiple processed images into the pixel interval evaluation model, and count the number of pixels in each pixel value interval of the multiple processed images by using the pixel interval evaluation model to obtain an image evaluation result output by the pixel interval evaluation model according to the statistical result.
In one embodiment, the evaluation model includes a custom evaluation model, at least one custom scheme is set in the custom evaluation model, and the image evaluation module is specifically configured to input the plurality of processing images into the custom evaluation model, and perform calculation according to the custom scheme and the plurality of processing images by using the custom evaluation model to obtain an image evaluation result output by the custom evaluation model.
In one embodiment, the evaluation model further comprises a performance evaluation model, and the apparatus further comprises:
the performance evaluation module is used for acquiring the running performance information of each image processing algorithm in the process of processing the reference data by a plurality of image processing algorithms; and analyzing the operation performance information of each image processing algorithm by using the performance evaluation model to obtain a performance evaluation result.
In one embodiment, the apparatus further comprises:
the report generating module is used for generating an evaluation report according to the image evaluation result and the performance evaluation result; the evaluation report includes an image evaluation result, a performance evaluation result, and a difference image generated from the plurality of processed images.
A computer device comprising a memory storing a computer program and a processor implementing the following steps when the computer program is executed:
acquiring reference data;
respectively processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein the versions of the plurality of image processing algorithms are different;
evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result; wherein the image evaluation results are used to characterize differences between the plurality of processed images.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring reference data;
respectively processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein the versions of the plurality of image processing algorithms are different;
evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result; wherein the image evaluation results are used to characterize differences between the plurality of processed images.
The terminal acquires the reference data according to the evaluation method and device of the image processing algorithm, the computer equipment and the storage medium; respectively processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; and evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result. In the embodiment of the disclosure, the evaluation model is distinguished from the image processing algorithm, that is, the evaluation code is distinguished from the algorithm code, so that the evaluation model is not affected when the image processing algorithm is iterated; the image processing algorithm is not affected when the evaluation model is modified. Compared with the prior art, the evaluation model and the image processing algorithm can be maintained independently, so that the maintenance difficulty is reduced, and the iteration efficiency of the image processing algorithm is improved.
Drawings
FIG. 1 is a diagram illustrating an exemplary embodiment of a method for evaluating an image processing algorithm;
FIG. 2 is a schematic flow chart diagram illustrating a method for evaluating an image processing algorithm in one embodiment;
FIG. 3a is a schematic illustration of image partitioning in one embodiment;
FIG. 3b is a schematic view of a region of interest in one embodiment;
FIG. 4 is a histogram of pixel count statistics in one embodiment;
FIG. 5 is a schematic flow chart of the operational performance evaluation step in one embodiment;
FIG. 6 is a diagram illustrating image data in reference data in one embodiment;
FIG. 7a is a diagram illustrating processing of image a according to one embodiment;
FIG. 7b is a schematic diagram of processing an image a' in one embodiment;
FIG. 7c is a diagram of a difference image in one embodiment;
FIG. 8 is a schematic view of a region of interest in another embodiment;
FIG. 9 is a block diagram showing an arrangement of an evaluation device of an image processing algorithm according to an embodiment;
FIG. 10 is a diagram showing an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The evaluation method of the image processing algorithm provided by the application can be applied to the application environment shown in fig. 1. The application environment may include a terminal 102 and a server 104, with the terminal 102 communicating with the server 104 over a network. The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the server 104 may be implemented by an independent server or a server cluster formed by a plurality of servers.
In one embodiment, as shown in fig. 2, there is provided an evaluation method of an image processing algorithm, which is described by taking the method as an example applied to the terminal in fig. 1, and includes the following steps:
in step 201, reference data is obtained.
Wherein the reference data comprises image data and/or parameter data. The parameter data may include image parameter data and algorithm parameter data, as shown in table 1, the image parameter data may include image width, image height, pixel bit depth, etc.; the algorithm parameter data comprises a T (filter kernel radius) threshold value, a Sigma color value, a Sigma space value, a filter kernel size, a transverse filter coefficient, a longitudinal filter coefficient, an edge point difference value type, a single-channel edge output image, a hysteresis threshold value, an operator aperture size, an image gradient amplitude sign, a binarization threshold value, a binarization maximum value, a threshold value type and the like of the algorithm. The embodiment of the present disclosure does not limit the parameter data.
TABLE 1
| Width of image | 453 |
| Height of image | 394 |
| |
8 |
| Image processing algorithm T threshold | 6 |
| Image processing algorithm |
7 |
| Image processing algorithm |
3 |
The terminal may acquire the image data in a plurality of ways, one of which is that the imaging device detects the detection object to obtain the image data, and then the terminal acquires the image data from the imaging device. For example, the ultrasound device performs an ultrasound scan on the detection object to obtain ultrasound image data, and then the terminal acquires the ultrasound image data from the ultrasound device.
In another method, after the imaging device detects the detection object, the obtained image data is stored in an image database, and then the terminal acquires the image data from the image database. For example, the ultrasound device performs ultrasound scanning on the detected object to obtain ultrasound image data, and stores the ultrasound image data in an image database; the terminal then acquires ultrasound image data from an image database.
The terminal may obtain the parameter data in various ways, and one way is that the terminal receives the parameter data input by the user. For example, the terminal receives parameter data such as a T threshold, a Sigma color value and the like of an image data algorithm input by a user.
In another mode, the terminal calculates parameter data according to the acquired image data and an image processing algorithm to be evaluated. For example, the terminal calculates parameter data such as image width and image height according to the acquired ultrasound image processing and image processing algorithm a and image processing algorithm B.
The embodiment of the disclosure does not limit the acquisition modes of the image data and the parameter data.
Wherein the versions of the plurality of image processing algorithms are different. For example, two bilateral filtering algorithms are preset, and the bilateral filtering algorithm a' is an iterative version of the bilateral filtering algorithm a. The embodiment of the present disclosure does not limit the version of the image processing algorithm.
The terminal may acquire a plurality of image processing algorithms from the server in advance and set the image processing algorithms in the terminal. When the image processing algorithms are evaluated, the terminal processes the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms.
For example, the terminal obtains the bilateral filtering algorithm a and the bilateral filtering algorithm a 'from the server in advance, and obtains the processed image a corresponding to the bilateral filtering algorithm a and the processed image a' corresponding to the bilateral filtering algorithm a 'after processing the reference data by using the bilateral filtering algorithm a and the bilateral filtering algorithm a'.
A plurality of image processing algorithms may be set in the server in advance. When the image processing algorithms are evaluated, the terminal sends the reference data to the server, and the reference data are processed by a plurality of image processing algorithms arranged in the server to obtain processing images corresponding to the image processing algorithms. The server then transmits the obtained processed image to the terminal, and the terminal receives the processed image corresponding to each image processing algorithm transmitted by the server.
And step 203, evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result.
Wherein the image evaluation results are used to characterize differences between the plurality of processed images. For example, the image evaluation result indicates the number of pixels having the same pixel value and the number of pixels having different pixel values in the processed image a and the processed image a'. The embodiment of the present disclosure does not limit the image evaluation result.
And after the terminal obtains the processing images corresponding to the image processing algorithms, determining the difference among the plurality of processing images by using the evaluation model to obtain an image evaluation result. It will be appreciated that the smaller the difference between the multiple processed images, the less the image is degraded by the iteration of the image processing algorithm.
And the evaluation model and the image processing algorithm are independent from each other, and the modification and the update of the evaluation model and the modification and the update of the image processing algorithm are not interfered with each other.
In the evaluation method of the image processing algorithm, a terminal acquires reference data; respectively processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; and evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result. In the embodiment of the disclosure, the evaluation model is distinguished from the image processing algorithm, that is, the evaluation code is distinguished from the algorithm code, so that the evaluation model is not affected when the image processing algorithm is iterated; the image processing algorithm is not affected when the evaluation model is modified. Compared with the prior art, the evaluation model and the image processing algorithm can be maintained independently, so that the maintenance difficulty is reduced, and the iteration efficiency of the image processing algorithm is improved. Furthermore, the evaluation model and the image processing algorithm are independent from each other, the evaluation model can be suitable for different image processing algorithms, and the effect of one-time development and reuse is realized.
In one embodiment, the evaluation model includes a global contrast evaluation model, and the step of evaluating the plurality of processed images using the evaluation model to obtain the image evaluation result may include: and inputting the plurality of processed images into a full-image comparison and evaluation model, and comparing the pixel value of each pixel point in the plurality of processed images by using the full-image comparison and evaluation model to obtain an image evaluation result output by the full-image comparison and evaluation model.
Illustratively, the terminal inputs the processed image a and the processed image a ' into a full-map contrast evaluation model, the full-map contrast evaluation model compares the pixel value of the pixel point 1 in the processed image a with the pixel value of the pixel point 1 in the processed image a ', and compares the pixel value of the pixel point 2 in the processed image a with the pixel value of the pixel point 2 in the processed image a '. And by analogy, comparing the pixel values of each pixel point of the processed image a and the processed image a', and then outputting an image evaluation result by the whole image comparison evaluation model.
The image evaluation result comprises the number and/or the positions of first pixel points with the same pixel value, the pixel difference value accords with the number and/or the positions of second pixel points with different preset thresholds, and the number of the second pixel points accounts for at least one of the proportion of the total number of the pixel points. As shown in table 2, the number of the first pixels with the same pixel value is 712558, and the number of the second pixels with different pixel values is 73874. The number of the second pixel points with the pixel difference value less than or equal to 1 is 12528, the number of the second pixel points with the pixel difference value less than or equal to 2 and 1 is 22523, and the number of the second pixel points with the pixel difference value less than or equal to 3 and 2 is 25552. In table 2, the preset Threshold value Tolerance Threshold includes 1, 2, and 3, and the preset Threshold value is not limited in the embodiment of the present disclosure.
TABLE 2
| Width of image | 1024 |
| Height of image | 768 |
| Total number of pixel points | 786432 |
| The number of pixel points with the same pixel value | 712558 |
| Number of pixels having different pixel values | 73874 |
| The number of pixel points with the pixel difference value less than or equal to 1 | 12528 |
| The number of pixel points with the pixel difference value of more than 1 and less than or equal to 2 | 22523 |
| The number of pixel points with the pixel difference value of more than 2 and less than or equal to 3 | 25552 |
| …… | |
| Number of pixels having different pixel values/total number of pixels | 9.39％ |
The full-image comparison evaluation model compares the pixel points of the processed image one by one, and can count and summarize the comparison result of each pixel point in the full-image processed image to form a statistical report.
In one embodiment, the step of evaluating the plurality of processed images by using the evaluation model to obtain the image evaluation result may include: and inputting the plurality of processing images into the partition comparison and evaluation model, and comparing the pixel values of the corresponding partitions in the plurality of processing images by using the partition comparison and evaluation model to obtain an image evaluation result output by the partition comparison and evaluation model.
After the terminal inputs the multiple processing images into the partition comparison and evaluation model, the partition comparison and evaluation model may perform partition processing on the multiple processing images, and set a preset threshold corresponding to each partition. As shown in fig. 3a, the preset threshold of the upper left corner region is 25, the partition comparison evaluation model compares the pixel value of each pixel in the upper left corner region of the processed image a with the pixel value of each pixel in the upper left corner region of the processed image a', and obtains the number and/or position of the pixels with the pixel difference value less than 25. And by analogy, comparing the processed image a with the processed image a' according to the preset threshold values respectively corresponding to the lower left corner region, the upper right corner region and the lower right corner region. Finally, the subarea comparison evaluation model outputs an image evaluation result according to the comparison results of the plurality of subareas
The partition comparison and evaluation model can also select an interested region from the multiple processed images and set a preset threshold corresponding to the interested region. As shown in fig. 3b, the preset threshold of the ROI is 5, the partition comparison and evaluation model compares the pixel value of each pixel in the ROI of the processed image a with the pixel value of each pixel in the ROI of the processed image a', and obtains the number and/or position of the pixels with the pixel difference smaller than 5.
As can be understood, the partition comparison and evaluation model performs regional comparison on the processed images, and may perform statistical summarization on the regional comparison results (especially the comparison results of the regions of interest) to form a statistical report.
In an embodiment, the step of obtaining the image evaluation result by evaluating the plurality of processed images using the evaluation model may include: inputting the multiple processed images into a pixel interval evaluation model, and utilizing the pixel interval evaluation model to count the number of pixel points in each pixel value interval in the multiple processed images to obtain an image evaluation result output by the pixel interval evaluation model according to the statistical result.
The pixel value interval evaluation model divides pixel values 0-255 into a plurality of pixel value intervals. For each processed image, the pixel interval evaluation model counts the number of pixel points in each pixel value interval. As shown in fig. 4, the number of pixels in each pixel value interval is shown by a histogram.
And a corresponding preset threshold value is also set for each pixel value interval in the pixel interval evaluation model. The pixel interval evaluation model compares corresponding pixel points of a plurality of processed images, and counts the number of the pixel points which are located in the same pixel value interval and the pixel difference value of which accords with a preset threshold value. As shown in fig. 4, a pixel value 0-30 is set as a pixel value interval, a preset threshold value of the pixel value interval is set as 20, the pixel interval evaluation model compares the processed image a with the corresponding pixel points in the processed image a', and counts the number of the pixel points whose pixel values are between 0 and 30 and whose pixel difference is less than 20.
Understandably, the pixel interval evaluation model performs statistics and summarization of a pixel value interval on each pixel point of the processed image to form a statistical form.
In one embodiment, the evaluation model includes a custom evaluation model having at least one custom scenario disposed therein. The step of evaluating the plurality of processed images by using the evaluation model to obtain the image evaluation result may include: and inputting the plurality of processing images into a user-defined evaluation model, and calculating by using the user-defined evaluation model according to the user-defined scheme and the plurality of processing images to obtain an image evaluation result output by the user-defined evaluation model.
The custom evaluation scheme may include at least one of a standard deviation formula, a root mean square formula, and a root mean square error formula.
The standard deviation formula is as in formula (1):
wherein S is standard deviation, n is the number of processed images, i is the serial number of pixel points, m is the total number of pixel points, xiIs the pixel value, x 'of the ith pixel point in one processed image'iThe pixel value of the ith pixel point in another processed image.
The root mean square formula is as formula (2):
wherein, XrmsIs the root mean square, x is the pixel value of the pixel point in the processed image, and N is the total number of the pixel points.
The root mean square error formula is as in formula (3):
wherein RMSE is root mean square error, n is the number of processed images, i is the serial number of pixel points, m is the total number of pixel points, wiIs a preset threshold value, x, of the ith pixel pointiIs the pixel value, x 'of the ith pixel point in one processed image'iThe pixel value of the ith pixel point in another processed image.
The terminal inputs the plurality of processed images into the user-defined evaluation model, the user-defined evaluation model carries out calculation according to a preset user-defined scheme, and image evaluation results such as the standard deviation of pixel values of the plurality of processed images, the root mean square of the pixel value of each processed image, the root mean square error of the pixel values of the plurality of processed images and the like can be calculated. And then, outputting the image evaluation result by the self-defined evaluation model.
The self-defined evaluation model can calculate the processed image according to a preset self-defined scheme and count and summarize the calculation result so as to form a statistical report.
In an embodiment, the evaluation model further includes a performance evaluation model, and on the basis of the above embodiment, as shown in fig. 5, the method may further include the following steps:
The operation performance information includes at least one of operation time, Central Processing Unit (CPU) occupancy, Graphics Processing Unit (GPU) occupancy, memory occupancy, and Input/Output (IO) performance.
And in the process of processing the reference data by the plurality of image processing algorithms, the terminal acquires the running performance information corresponding to each image processing algorithm.
And step 302, analyzing the operation performance information of each image processing algorithm by using a performance evaluation model to obtain a performance evaluation result.
After the terminal obtains the operation performance information corresponding to each image processing algorithm, the performance evaluation model compares and analyzes the operation performance information of the image processing algorithms to obtain a performance evaluation result.
In the embodiment, the terminal acquires the running performance information of each image processing algorithm in the process of processing the reference data by a plurality of image processing algorithms; and analyzing the operation performance information of each image processing algorithm by using the performance evaluation model to obtain a performance evaluation result. In the embodiment of the disclosure, the terminal can analyze and compare the operation performance of the image processing algorithm, and on the basis of ensuring that the image is not degraded, a programmer can know which image processing algorithm has better operation performance, so that the iteration efficiency of the image processing algorithm is improved.
In one embodiment, the evaluation model may include one or more of a global contrast evaluation model, a partition contrast evaluation model, a pixel interval evaluation model, a custom evaluation model, a performance evaluation model. In practical applications, other evaluation models besides the above evaluation model may also be used, which is not limited in the embodiments of the present disclosure.
It will be appreciated that if the evaluation model includes multiple evaluation models, the processed image may be evaluated from multiple dimensions, thereby making the evaluation more comprehensive.
In one embodiment, on the basis of the above embodiment, the method may further include: and generating an evaluation report according to the image evaluation result and the performance evaluation result. Wherein the evaluation report includes the image evaluation result, the performance evaluation result, and a difference image generated from the plurality of processed images.
Taking the image data shown in fig. 6 and the parameter data shown in table 3 as reference data, and the image processing algorithms to be evaluated are the bilateral filtering algorithm a and the bilateral filtering algorithm a ', as an example, the generated evaluation report may include the processed image a shown in fig. 7a, the processed image a' shown in fig. 7b, and the difference image shown in fig. 7 c.
TABLE 3
| Width of image | 453 |
| Height of image | 394 |
| |
8 |
| Bilateral filtering algorithm T threshold | 6 |
| Bilateral filter algorithm |
7 |
| Bilateral filter algorithm Sigma |
3 |
The evaluation report may further include the image evaluation result output by the full-image comparison evaluation model, as shown in table 4, the total number of pixels is 178482, the number of pixels with the same pixel value is 128168, the number of pixels with different pixel values is 50314, and the ratio of the number of pixels with different pixel values to the total number of pixels is 28.19%:
TABLE 4
| Width of image | 453 |
| Height of image | 394 |
| Total number of pixel points | 178482 |
| The number of pixel points with the same pixel value | 128168 |
| Number of pixels having different pixel values | 50314 |
| The number of pixel points with the pixel difference value less than or equal to 1 | 49913 |
| The number of pixel points with the pixel difference value of more than 1 and less than or equal to 2 | 399 |
| The number of pixel points with the pixel difference value of more than 2 and less than or equal to 3 | 2 |
| Number of pixels with pixel difference value > 3 | 0 |
| Number of pixels having different pixel values/total number of pixels | 28.19％ |
As shown in fig. 8 and table 5, the total number of pixels is 124992, the number of pixels with the same pixel value is 114703, the number of pixels with different pixel values is 10289, and the ratio of the number of pixels with different pixel values to the total number of pixels is 8.23%:
TABLE 5
| Region of interest size (x, y) | 40，30 |
| Region of interest size (w, h) | 372，336 |
| Preset threshold value of interested area | 1 |
| Total number of pixels | 124992 |
| The number of pixel points with the same pixel value | 114703 |
| Number of pixels having different pixel values | 10289 |
| The number of pixel points with the pixel difference value less than or equal to 2 | 10289 |
| Number of pixels with pixel difference value > 2 | 0 |
| Number of pixels having different pixel values/total number of pixels | 8.23％ |
The evaluation report may further include an image evaluation result output by the custom evaluation model, as shown in table 6, where the total number of pixels is 178482, the number of pixels with the same pixel value is 178480, the number of pixels with different pixel values is 2, and the ratio of the number of pixels with different pixel values to the total number of pixels is about 0%:
TABLE 6
The self-defined scheme set in the self-defined evaluation model is a root mean square error formula.
After the terminal obtains the image evaluation result, an evaluation report can be generated according to the image evaluation result. It can be understood that, if the ratio of the number of the pixel points with different pixel values to the total number of the pixel points of the processed images processed by different image processing algorithms is smaller, it indicates that the difference between the plurality of processed images is smaller, and the processing effects of the plurality of image processing algorithms are consistent. On the contrary, if the ratio is larger, it indicates that the difference between the plurality of processed images is larger, the processing effect difference of the plurality of image processing algorithms is larger, and a developer can correspondingly improve the image processing algorithms according to the evaluation report.
Further, the evaluation report may also include performance evaluation results, as shown in table 7:
TABLE 7
| Bilateral filtering model A | Bilateral filter model A' | |
| Run time | 0.021ms | 0.026ms |
| CUP occupancy rate | 0.29％ | 0.34％ |
| GPU occupancy | 0.00％ | 0.00％ |
| Memory occupancy rate | 0.37MB | 0.37MB |
| IO | 0.00％ | 0.00％ |
In one embodiment, the method may further include: and adjusting the preset threshold value, and utilizing the evaluation model to evaluate the plurality of processed images again to obtain a new image evaluation result.
The terminal can adjust preset thresholds in the full-image comparison evaluation model, the partition comparison evaluation model, the pixel interval evaluation model and the user-defined evaluation model, and then evaluates the processed image again according to the adjusted preset thresholds by using the full-image comparison evaluation model, the partition comparison evaluation model, the pixel interval evaluation model and the user-defined evaluation model to obtain a new image evaluation result.
In the embodiment, the evaluation report can clearly and intuitively show the comparison results and the statistical results of multiple dimensions, and the evaluation result is more objective.
It should be understood that although the various steps in the flowcharts of fig. 2-5 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not limited to being performed in the exact order illustrated and, unless explicitly stated herein, may be performed in other orders. Moreover, at least some of the steps in fig. 2-5 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps or stages.
In one embodiment, as shown in fig. 9, there is provided an evaluation apparatus of an image processing algorithm, including:
a data obtaining module 401, configured to obtain reference data;
an image processing module 402, configured to process the reference data by using a plurality of preset image processing algorithms, respectively, to obtain processing images corresponding to the image processing algorithms; wherein the versions of the plurality of image processing algorithms are different;
an image evaluation module 403, configured to evaluate the multiple processed images by using the evaluation model to obtain an image evaluation result; wherein the image evaluation results are used to characterize differences between the plurality of processed images.
In one embodiment, the evaluation model includes a full-image comparison evaluation model, and the image evaluation module 403 is specifically configured to input the multiple processed images into the full-image comparison evaluation model, and compare the pixel values of each pixel point in the multiple processed images by using the full-image comparison evaluation model to obtain an image evaluation result output by the full-image comparison evaluation model.
In one embodiment, the evaluation model includes a partition comparison evaluation model, and the image evaluation module 403 is specifically configured to input the multiple processing images into the partition comparison evaluation model, and compare pixel values of corresponding partitions in the multiple processing images with the partition comparison evaluation model to obtain an image evaluation result output by the partition comparison evaluation model.
In one embodiment, the evaluation model includes a pixel interval evaluation model, and the image evaluation module 403 is specifically configured to input the multiple processing images into the pixel interval evaluation model, and count the number of pixels in each pixel value interval of the multiple processing images by using the pixel interval evaluation model to obtain an image evaluation result output by the pixel interval evaluation model according to the statistical result.
In one embodiment, the evaluation model includes a custom evaluation model, at least one custom scheme is set in the custom evaluation model, and the image evaluation module 403 is specifically configured to input the multiple processed images into the custom evaluation model, and perform calculation according to the custom scheme and the multiple processed images by using the custom evaluation model to obtain an image evaluation result output by the custom evaluation model.
In one embodiment, the evaluation model further comprises a performance evaluation model, and the apparatus further comprises:
the performance evaluation module is used for acquiring the running performance information of each image processing algorithm in the process of processing the reference data by a plurality of image processing algorithms; and analyzing the operation performance information of each image processing algorithm by using the performance evaluation model to obtain a performance evaluation result.
In one embodiment, the apparatus further comprises:
the report generation module is used for generating an evaluation report according to the image evaluation result and the performance evaluation result; the evaluation report includes an image evaluation result, a performance evaluation result, and a difference image generated from the plurality of processed images.
The specific definition of the evaluation device for the image processing algorithm can be referred to the above definition of the evaluation method for the image processing algorithm, and is not described herein again. The various modules in the evaluation means of the above-described image processing algorithm may be implemented in whole or in part by software, hardware, and combinations thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 10. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a method of evaluating an image processing algorithm. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 10 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring reference data;
respectively processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein the versions of the plurality of image processing algorithms are different;
evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result; wherein the image evaluation result is used to characterize the differences between the plurality of processed images.
In one embodiment, the evaluation model comprises a full-map contrast evaluation model, and the processor when executing the computer program further performs the steps of:
and inputting the plurality of processed images into a full-image comparison and evaluation model, and comparing the pixel value of each pixel point in the plurality of processed images by using the full-image comparison and evaluation model to obtain an image evaluation result output by the full-image comparison and evaluation model.
In one embodiment, the evaluation model comprises a partition comparison evaluation model, and the processor when executing the computer program further performs the steps of:
and inputting the plurality of processing images into the partition comparison and evaluation model, and comparing the pixel values of the corresponding partitions in the plurality of processing images by using the partition comparison and evaluation model to obtain an image evaluation result output by the partition comparison and evaluation model.
In one embodiment, the evaluation model comprises a pixel interval evaluation model, the pixel interval evaluation model is provided with a plurality of pixel value intervals, and the processor when executing the computer program further implements the following steps:
inputting the multiple processed images into a pixel interval evaluation model, and utilizing the pixel interval evaluation model to count the number of pixel points in each pixel value interval in the multiple processed images to obtain an image evaluation result output by the pixel interval evaluation model according to the statistical result.
In one embodiment, the evaluation model includes a custom evaluation model, at least one custom solution is set in the custom evaluation model, and the processor executes the computer program to further implement the following steps:
and inputting the plurality of processing images into a user-defined evaluation model, and calculating by using the user-defined evaluation model according to the user-defined scheme and the plurality of processing images to obtain an image evaluation result output by the user-defined evaluation model.
In one embodiment, the evaluation model further comprises a performance evaluation model, and the processor when executing the computer program further performs the steps of:
acquiring the running performance information of each image processing algorithm in the process of processing the reference data by a plurality of image processing algorithms;
and analyzing the operation performance information of each image processing algorithm by using the performance evaluation model to obtain a performance evaluation result.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
generating an evaluation report according to the image evaluation result and the performance evaluation result; the evaluation report includes an image evaluation result, a performance evaluation result, and a difference image generated from the plurality of processed images.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring reference data;
respectively processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein the versions of the plurality of image processing algorithms are different;
evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result; wherein the image evaluation results are used to characterize differences between the plurality of processed images.
In one embodiment, the evaluation model comprises a full-map contrast evaluation model, the computer program when executed by the processor further performing the steps of:
and inputting the plurality of processed images into a full-image comparison and evaluation model, and comparing the pixel value of each pixel point in the plurality of processed images by using the full-image comparison and evaluation model to obtain an image evaluation result output by the full-image comparison and evaluation model.
In one embodiment, the evaluation model comprises a partition comparison evaluation model, and the computer program when executed by the processor further performs the steps of:
and inputting the plurality of processing images into the partition comparison and evaluation model, and comparing the pixel values of the corresponding partitions in the plurality of processing images by using the partition comparison and evaluation model to obtain an image evaluation result output by the partition comparison and evaluation model.
In one embodiment, the evaluation model comprises a pixel interval evaluation model, the pixel interval evaluation model being provided with a plurality of pixel value intervals, the computer program when executed by the processor further realizing the steps of:
inputting the multiple processed images into a pixel interval evaluation model, and utilizing the pixel interval evaluation model to count the number of pixel points in each pixel value interval in the multiple processed images to obtain an image evaluation result output by the pixel interval evaluation model according to the statistical result.
In one embodiment, the evaluation model includes a custom evaluation model having at least one custom scenario set therein, and the computer program when executed by the processor further performs the steps of:
and inputting the plurality of processing images into a user-defined evaluation model, and calculating by using the user-defined evaluation model according to the user-defined scheme and the plurality of processing images to obtain an image evaluation result output by the user-defined evaluation model.
In one embodiment, the evaluation model further comprises a performance evaluation model, the computer program when executed by the processor further implementing the steps of:
acquiring the running performance information of each image processing algorithm in the process of processing the reference data by a plurality of image processing algorithms;
and analyzing the operation performance information of each image processing algorithm by using the performance evaluation model to obtain a performance evaluation result.
In one embodiment, the computer program when executed by the processor further performs the steps of:
generating an evaluation report according to the image evaluation result and the performance evaluation result; the evaluation report includes an image evaluation result, a performance evaluation result, and a difference image generated from the plurality of processed images.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is specific and detailed, but not to be understood as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method of evaluating an image processing algorithm, the method comprising:
acquiring reference data;
processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein versions of the plurality of image processing algorithms are different;
evaluating the plurality of processed images by utilizing an evaluation model to obtain an image evaluation result; wherein the image evaluation result is used to characterize differences between the plurality of processed images.
2. The method of claim 1, wherein the evaluation model comprises a global contrast evaluation model, and wherein evaluating the plurality of processed images using the evaluation model to obtain image evaluation results comprises:
and inputting the plurality of processed images into the full-image comparison and evaluation model, and comparing the pixel value of each pixel point in the plurality of processed images by using the full-image comparison and evaluation model to obtain the image evaluation result output by the full-image comparison and evaluation model.
3. The method of claim 1, wherein the evaluation model comprises a partition contrast evaluation model, and wherein evaluating the plurality of processed images using the evaluation model to obtain image evaluation results comprises:
inputting a plurality of processing images into the subarea comparison and evaluation model, and comparing pixel values of corresponding subareas in the plurality of processing images by using the subarea comparison and evaluation model to obtain the image evaluation result output by the subarea comparison and evaluation model.
4. The method according to claim 1, wherein the evaluation model comprises a pixel interval evaluation model, the pixel interval evaluation model is provided with a plurality of pixel value intervals, and the evaluation of the plurality of processed images by the evaluation model to obtain an image evaluation result comprises:
inputting a plurality of processed images into the pixel interval evaluation model, and counting the number of pixel points in each pixel value interval in the plurality of processed images by using the pixel interval evaluation model to obtain the image evaluation result output by the pixel interval evaluation model according to the statistical result.
5. The method of claim 1, wherein the evaluation model comprises a custom evaluation model, at least one custom solution is set in the custom evaluation model, and the evaluating the plurality of processed images by using the evaluation model to obtain an image evaluation result comprises:
and inputting a plurality of processing images into the user-defined evaluation model, and calculating according to the user-defined scheme and the plurality of processing images by using the user-defined evaluation model to obtain the image evaluation result output by the user-defined evaluation model.
6. The method of any of claims 1-5, wherein the evaluation model further comprises a performance evaluation model, the method further comprising:
acquiring the running performance information of each image processing algorithm in the process of processing the reference data by the plurality of image processing algorithms;
and analyzing the operation performance information of each image processing algorithm by using the performance evaluation model to obtain a performance evaluation result.
7. The method of claim 6, wherein after analyzing the operational performance information of each of the image processing algorithms using the performance evaluation model to obtain a performance evaluation result, the method further comprises:
generating an evaluation report according to the image evaluation result and the performance evaluation result; the evaluation report includes the image evaluation result, the performance evaluation result, and a difference image generated from a plurality of the processed images.
8. An apparatus for evaluating an image processing algorithm, the apparatus comprising:
the data acquisition module is used for acquiring reference data;
the image processing module is used for processing the reference data by utilizing a plurality of preset image processing algorithms to obtain processing images corresponding to the image processing algorithms; wherein versions of the plurality of image processing algorithms are different;
the image evaluation module is used for evaluating the plurality of processed images by using an evaluation model to obtain an image evaluation result; wherein the image evaluation result is used to characterize differences between the plurality of processed images.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011401108.3A CN114612367A (en) | 2020-12-04 | 2020-12-04 | Evaluation method and device of image processing algorithm, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011401108.3A CN114612367A (en) | 2020-12-04 | 2020-12-04 | Evaluation method and device of image processing algorithm, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114612367A true CN114612367A (en) | 2022-06-10 |
Family
ID=81856817
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011401108.3A Pending CN114612367A (en) | 2020-12-04 | 2020-12-04 | Evaluation method and device of image processing algorithm, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114612367A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115809188A (en) * | 2022-08-03 | 2023-03-17 | 宁德时代新能源科技股份有限公司 | Debugging method, device, equipment, medium and program product of image detection algorithm |
| CN117455903A (en) * | 2023-12-18 | 2024-01-26 | 深圳市焕想科技有限公司 | Sports apparatus state evaluation method based on image processing technology |
-
2020
- 2020-12-04 CN CN202011401108.3A patent/CN114612367A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115809188A (en) * | 2022-08-03 | 2023-03-17 | 宁德时代新能源科技股份有限公司 | Debugging method, device, equipment, medium and program product of image detection algorithm |
| CN115809188B (en) * | 2022-08-03 | 2024-02-06 | 宁德时代新能源科技股份有限公司 | Image detection algorithm debugging method, device, equipment and storage medium |
| CN117455903A (en) * | 2023-12-18 | 2024-01-26 | 深圳市焕想科技有限公司 | Sports apparatus state evaluation method based on image processing technology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tao et al. | Manifold ranking-based matrix factorization for saliency detection | |
| CN109086711B (en) | Face feature analysis method and device, computer equipment and storage medium | |
| US9342870B2 (en) | Tree-based linear regression for denoising | |
| CN109377500B (en) | Image segmentation method based on neural network and terminal equipment | |
| CN109711508B (en) | Image processing method and device | |
| CN110533632B (en) | Image blurring tampering detection method and device, computer equipment and storage medium | |
| CN110992243B (en) | Intervertebral disc cross-section image construction method, device, computer equipment and storage medium | |
| CN114612367A (en) | Evaluation method and device of image processing algorithm, computer equipment and storage medium | |
| US20220285024A1 (en) | Facilitating interpretability of classification model | |
| CN112132265A (en) | Model training method, cup-to-tray ratio determining method, device, equipment and storage medium | |
| CN111652862A (en) | Spine alignment classification method, computer device and storage medium | |
| CN113132560A (en) | Video processing method, computer equipment and computer readable storage medium | |
| CN110914810B (en) | Monkey testing method and terminal | |
| US20140355959A1 (en) | Multi-frame patch correspondence identification in video | |
| CN111489318B (en) | Medical image enhancement method and computer-readable storage medium | |
| US10997715B2 (en) | System and method for implementing a medical imaging composite average and clustering technique | |
| CN115731442A (en) | Image processing method, image processing device, computer equipment and storage medium | |
| CN115147296A (en) | Hyperspectral image correction method, device, computer equipment and storage medium | |
| CN114332297A (en) | Image drawing method and device, computer equipment and storage medium | |
| CN113158858A (en) | Behavior analysis method and system based on deep learning | |
| US9704056B2 (en) | Computing hierarchical computations for computer vision calculations | |
| US20230316700A1 (en) | Non-transitory computer readable recording medium, visualization method and information processing apparatus | |
| Chen et al. | Speckle noise removal by energy models with new regularization setting | |
| TWI622959B (en) | A medical image processing system and its method | |
| Zhang | Novel Statistical Learning Methods for High-Dimensional Complex Biomedical Data Analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |