CN115908394A - Tiny fracture detection method and system based on target detection and graph attention machine mechanism - Google Patents
Tiny fracture detection method and system based on target detection and graph attention machine mechanism Download PDFInfo
- Publication number
- CN115908394A CN115908394A CN202211710001.6A CN202211710001A CN115908394A CN 115908394 A CN115908394 A CN 115908394A CN 202211710001 A CN202211710001 A CN 202211710001A CN 115908394 A CN115908394 A CN 115908394A
- Authority
- CN
- China
- Prior art keywords
- image
- fracture
- chest
- detection
- fine
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a method and a system for detecting tiny fracture based on target detection and an image attention machine mechanism, wherein the method comprises the steps of obtaining a plurality of time sequence chest CT images, preprocessing the images, and constructing a training sample set by the chest CT images marked with tiny fracture areas; training a fine fracture detection model by using a training sample set, extracting image characteristics by using a convolutional neural network, extracting a graph attention matrix by using the graph convolutional neural network, combining the image characteristics and the graph attention matrix to obtain final characteristics, and outputting a fine fracture region prediction frame by combining a YOLO target detection network based on the final characteristics; and inputting the chest CT image to be detected into the trained fine fracture detection model, outputting a fine fracture region and a fracture type, optimizing and identifying the result by adopting a target tracking algorithm, improving the accuracy of fine fracture region identification, and realizing accurate positioning of the fine fracture region in the chest CT image.
Description
Technical Field
The invention relates to the technical field of medical image processing, in particular to a method and a system for detecting fine fractures based on target detection and a drawing and attention mechanism.
Background
The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.
When the chest of a human body is attacked by external pressure or violence, tiny fractures often occur at the hit part, the fractured ends are broken inwards, and internal organs of the chest can be damaged, so the chest fractures have high morbidity and certain death risk. Among them, rib fracture is the most common injury in chest tiny fracture, and is also one of the main contents of medical identification and judicial identification.
At present, computer Tomography (CT) is a main method for diagnosing fine chest fractures, and physicians usually position the fine chest fractures and complications thereof according to conventional CT chest scanning images, but because the chest part is large and contains various bones such as ribs, clavicles, sternum, vertebrae and the like, the fine fractures are usually hidden and difficult to find, and the CT chest scanning images generally comprise dozens of to hundreds of images, the positioning of the fracture position in each CT chest image is a mechanical and tedious task, and if the number of patients is large, the workload for detecting the fine chest fractures is very complicated and huge. Meanwhile, the inevitable visual fatigue is considered to exist in the manual detection and identification, so that missed diagnosis and misdiagnosis appear occasionally, and the accuracy for detecting the fine fracture of the chest is low.
With the rapid development of the application of the deep learning technology in medical imaging, for example, the detection of lesions such as lung nodules, breast fractures, breast molybdenum targets and the like has more research and plays a higher role in practical application. However, in the prior art, the method for automatically identifying the fracture position of the chest detects the CT images of the chest one by one, ignores the time sequence relationship between the images, and ignores the position information of the image features in the images during the detection process, and the position information is particularly important for positioning a small target, thereby resulting in low accuracy of positioning the fracture position.
Disclosure of Invention
In order to solve the defects of the prior art, the invention provides a method and a system for detecting the tiny fracture based on target detection and an image attention machine mechanism.
In a first aspect, the present disclosure provides a method for detecting a fine fracture based on target detection and a graph attention machine, comprising:
acquiring a plurality of time sequence chest CT images, preprocessing the images, and constructing a training sample set by taking the chest CT images marked with small fracture areas as training samples;
training a tiny fracture detection model by using a training sample set; the training process comprises the following steps: inputting a training sample image into a fine fracture detection model, extracting image characteristics by using a convolutional neural network, extracting a graph attention matrix by using the graph convolutional neural network, combining the image characteristics and the graph attention matrix to obtain final characteristics, outputting a fine fracture region prediction frame by combining a YOLO target detection network based on the final characteristics, and training the fine fracture detection model according to the distance loss between the prediction frame and the target frame;
and inputting the chest CT image to be detected into the trained tiny fracture detection model, and outputting a tiny fracture area and a fracture type.
In a further technical scheme, the pretreatment comprises:
stacking a plurality of time sequence chest CT images to synthesize a multi-channel image;
and resampling according to the inter-pixel distance of the chest CT image to finish normalization.
According to a further technical scheme, the image feature and the graph attention matrix are combined to obtain a final feature, and the method comprises the following steps:
calculating a graph attention matrix extracted by a graph convolution neural network by using 1 multiplied by 1 convolution, and changing a final nonlinear activation function into sigmoid to enable a value range of a calculation result to be between 0 and 1;
and multiplying the calculated graph attention force matrix by the characteristic graph extracted by the convolutional neural network to obtain the final characteristic.
According to the further technical scheme, after a time sequence chest CT image is input into a tiny fracture detection model, corresponding target detection prediction frames are sequentially output, an output result is input into a tiny fracture detection optimization model, and the post-processing strengthening is carried out on the whole image sequence prediction result by adopting a Z-direction-based NMS algorithm and a target tracking algorithm.
The further technical scheme adopts an NMS algorithm and a target tracking algorithm based on the z direction, and comprises the following steps:
mapping the position of the prediction frame with the highest confidence level in the previous chest CT image and the position of the prediction frame with the highest confidence level in the next chest CT image into the current chest CT image, taking the position of the prediction frame with the highest confidence level in the current chest CT image as a central point, and comparing the position with the mapping position;
and if the z coordinate of the mapping position is within a preset value range and the IOU of the prediction frame of the mapping position and the prediction frame of the central point position is larger than a set threshold value, deleting the prediction frame of the mapping position, and outputting the prediction frame with the highest confidence level in the current chest CT image as a final prediction frame, otherwise, outputting the prediction frame of the mapping position as the final prediction frame.
According to a further technical scheme, the value range of the z coordinate is determined according to the distance between pixels of the chest CT image.
In a second aspect, the present disclosure provides a system for detecting a fine fracture based on target detection and a graphical attention machine, comprising:
the training sample construction module is used for acquiring a plurality of time-sequence chest CT images, preprocessing the images, and constructing a training sample set by taking the chest CT images marked with the tiny fracture areas as training samples;
the fine fracture detection model training module is used for training a fine fracture detection model by utilizing a training sample set; the training process comprises the following steps: inputting a training sample image into a fine fracture detection model, extracting image characteristics by using a convolutional neural network, extracting a graph attention matrix by using the graph convolutional neural network, combining the image characteristics and the graph attention matrix to obtain final characteristics, outputting a fine fracture region prediction frame by combining a YOLO target detection network based on the final characteristics, and training the fine fracture detection model according to the distance loss between the prediction frame and the target frame;
and the tiny fracture detection module is used for inputting the chest CT image to be detected into a tiny fracture detection model after training and outputting a tiny fracture area and a fracture type.
The technical scheme further comprises a tiny fracture detection optimization module used for inputting the output result into a tiny fracture detection optimization model, and performing post-processing reinforcement on the whole image sequence prediction result by adopting a Z-direction-based NMS algorithm and a target tracking algorithm.
In a third aspect, the present disclosure also provides an electronic device comprising a memory and a processor, and computer instructions stored on the memory and executed on the processor, wherein the computer instructions, when executed by the processor, perform the steps of the method of the first aspect.
In a fourth aspect, the present disclosure also provides a computer-readable storage medium for storing computer instructions which, when executed by a processor, perform the steps of the method of the first aspect.
The above one or more technical solutions have the following beneficial effects:
1. the invention provides a method and a system for detecting tiny fracture based on target detection and an image attention machine mechanism.
2. The invention also adopts a Z-direction-based NMS algorithm to perform Z-direction NMS fusion on the detection result of each CT image and filter the prediction result, thereby obtaining a more concise tiny fracture detection result.
3. The method effectively solves the problem of inaccurate positioning of the tiny fracture in the prediction result, particularly small targets with fracture areas smaller than 32 multiplied by 32 pixels, and compared with the traditional image detection, the method can effectively reduce the time consumption of the tiny fracture prediction, shorten the analysis time and provide auxiliary diagnosis results as early as possible.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
FIG. 1 is a flow chart of a method for detecting a small fracture based on target detection and a graph attention machine according to an embodiment of the present invention;
FIG. 2 is a schematic view of a fusion of CT images of the breast in accordance with an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a thin fracture detection model according to an embodiment of the present invention;
FIG. 4 is a graph showing the results of the detection of a small fracture area in the first embodiment of the present invention.
Detailed Description
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an", and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
Example one
Aiming at the problem of low accuracy of identification and detection of a fine fracture region in a chest CT image in the prior art, the embodiment provides a fine fracture detection method based on target detection and a drawing attention mechanism, as shown in fig. 1, including the following steps:
s1, acquiring a plurality of time sequence chest CT images, preprocessing the images, and constructing a training sample set by taking the chest CT images marked with small fracture areas as training samples.
And S2, training a tiny fracture detection model by using the training sample set. The training process is as follows: inputting a training sample image into a fine fracture detection model, extracting image features by using a convolutional neural network, extracting a graph attention matrix through the graph convolutional neural network, combining the image features and the graph attention matrix to obtain final features, outputting a fine fracture area prediction frame by combining a YOLO target detection network based on the final features, and training the fine fracture detection model according to the distance loss between the prediction frame and the target frame.
And S3, inputting the chest CT image to be detected into the trained tiny fracture detection model, and outputting a tiny fracture area and a fracture type.
In the step S1, a plurality of time series chest CT images are acquired, and the images are preprocessed, where the image preprocessing includes: and stacking the time-sequence multiple images to synthesize a multi-channel image. In this embodiment, adjacent 3 images are stacked to form a 3-channel image, as shown in fig. 2, after the breast CT image is pre-trained, a short time sequence feature can be added to the original image data, so as to optimize the detection result.
The preprocessing further includes resampling according to spacing, i.e., inter-pixel distance, of the CT image of the breast, where the spacing refers to an actual distance represented between two adjacent pixels in the image, for example, 1 pixel represents a distance of 0.5 mm in the chest of a human body, and resampling the CT image of the breast by using a clip.
Meanwhile, in this embodiment, the small fracture area in the chest CT image is labeled by using the target frame, and the fracture type of the small fracture area is labeled, that is, the rib label, the clavicle label, the sternum label and the vertebra label are labeled, and the chest CT image labeled with the small fracture area is used as a training sample, so as to construct a training sample set.
In the step S2, the fine fracture detection model is trained using the training sample set. As shown in fig. 2, in the training process, firstly, a training sample image is input into a fine fracture detection model, image features are extracted by using a convolutional neural network, an image attention matrix is extracted by using the convolutional neural network, and the image features and the image attention matrix are combined to obtain final features.
Specifically, as shown in fig. 3, for each input chest CT image, firstly, the convolutional neural network CNN and the graph convolutional neural network are used to extract an image feature and a graph attention matrix, respectively, and the extracted image feature is processed through the graph attention matrix, so as to implement a simulated attention mechanism and improve the accuracy of subsequent target detection. In order to facilitate the fusion of the attention force matrix and the image features, in this embodiment, a 1 × 1 convolution is further used to calculate the attention force matrix (actually equivalent to the image features) extracted by the graph convolution neural network, the last nonlinear activation function is changed to sigmoid, the value range of the calculation result is between 0 and 1, and the output image features are consistent with the image features extracted by the CNN in dimension through this calculation mode.
And then, multiplying the calculated graph convolution neural network characteristic serving as an attention mechanism by a characteristic graph extracted by the CNN to obtain a final characteristic, namely performing dot multiplication operation on the characteristic graph extracted by the CNN to simulate the attention mechanism, adding a calculation result of the attention mechanism into a characteristic fusion layer of an original algorithm, and enhancing small target position information lost due to downsampling in a subsequent YOLO target detection network so as to achieve the purpose of optimizing detection.
And then, based on the final characteristics, combining a YOLO target detection network, outputting a small fracture area prediction frame, and training a small fracture detection model according to the distance loss between the prediction frame and the target frame. The YOLO target detection network is an existing target detection network and is used for recognizing and detecting a target in an image.
And finally, in the step S3, inputting the chest CT image of the time sequence to be detected into the trained tiny fracture detection model, and outputting a tiny fracture area and a fracture type.
Furthermore, in order to solve the problem of not considering the time sequence relationship between images pointed out in the background art, the embodiment further provides a fine fracture detection optimization model on the basis of the fine fracture detection model, and the z-direction-based NMS algorithm and the target tracking algorithm are adopted to realize post-processing reinforcement on the prediction result of the whole image sequence, optimize the prediction result of adjacent images and improve the detection rate.
The NMS algorithm is a non-maximum value inhibition algorithm, the z direction is a depth direction, namely the sequential direction of an input time sequence chest CT image, an NMS algorithm and a target tracking algorithm based on the z direction are adopted, specifically, after the time sequence chest CT image is input into a tiny fracture detection model, corresponding prediction frames for target detection are sequentially output, the prediction frame position with the highest reliability in the last chest CT image (namely the probability value of a certain fracture type in a fracture region in the chest CT image is detected through the model) and the prediction frame position with the highest reliability in the next chest CT image are mapped into the current chest CT image through the target tracking algorithm, then the prediction frame is filtered through the NMS algorithm based on the z direction, namely the prediction frame position with the highest reliability in the current chest CT image is used as a central point, the position and the mapping position are compared, if the z coordinate of the mapping position is within a preset value range (namely within an inhibition range), and the IOU of the prediction frame of the mapping position and the prediction frame of the central point position is larger than a set threshold value, the prediction frame of the current chest CT image is deleted, otherwise, the prediction frame output is used as a final prediction frame, and the final prediction scheme of the prediction frame is output neatly, and the prediction frame is output, and the target prediction frame is output. In addition, the value range of the z coordinate and the set threshold are preset values, wherein the value range of the z coordinate is determined according to spacing of the chest CT image.
And finally, executing the step S3, inputting the chest CT image to be detected into the trained tiny fracture detection model, and outputting the tiny fracture area and the fracture type of the chest CT image to be detected through the tiny fracture detection optimization model, as shown in fig. 4.
In summary, in the embodiment, the image in the european space is mapped to the non-european space by mapping the space, so as to obtain the graph attention force matrix by using the graph convolution neural network, combine the features extracted by the CNN with the graph attention force matrix to obtain the result after each pixel point in the feature map is weighted, and detect the fine fracture region in each chest CT image; then, the output tiny fracture area is used as input, and a target tracking algorithm is used for post-processing the result, so that the recall rate is further improved; and finally, performing Z-direction NMS fusion on each CT image detection result, filtering the result to obtain a simpler tiny fracture detection result, and storing the final detection result as a json-format file, so that visual auxiliary diagnosis results can be provided for doctors conveniently.
The method effectively solves the problem of inaccurate positioning of the tiny fracture in the prediction result, particularly small targets with fracture areas smaller than 32 multiplied by 32 pixels, and compared with the traditional image detection, the method can effectively reduce the time consumption of the tiny fracture prediction, shorten the analysis time and provide auxiliary diagnosis results as early as possible.
Example two
The embodiment provides a tiny fracture detection system based on target detection and a graph attention machine mechanism, which comprises:
the training sample construction module is used for acquiring a plurality of time sequence chest CT images, preprocessing the images, and constructing a training sample set by taking the chest CT images marked with small fracture areas as training samples;
the tiny fracture detection model training module is used for training a tiny fracture detection model by utilizing a training sample set; the training process comprises the following steps: inputting a training sample image into a fine fracture detection model, extracting image characteristics by using a convolutional neural network, extracting a graph attention matrix by using the graph convolutional neural network, combining the image characteristics and the graph attention matrix to obtain final characteristics, outputting a fine fracture region prediction frame by combining a YOLO target detection network based on the final characteristics, and training the fine fracture detection model according to the distance loss between the prediction frame and the target frame;
and the tiny fracture detection module is used for inputting the chest CT image to be detected into a tiny fracture detection model after training and outputting a tiny fracture area and a fracture type.
Further comprising: and the tiny fracture detection optimization module is used for inputting the output result into a tiny fracture detection optimization model, and performing post-processing reinforcement on the whole image sequence prediction result by adopting a z-direction-based NMS algorithm and a target tracking algorithm.
EXAMPLE III
The present embodiment provides an electronic device comprising a memory and a processor, and computer instructions stored on the memory and executed on the processor, wherein the computer instructions, when executed by the processor, perform the steps of the method for detecting a fine fracture based on object detection and a graph attention mechanism as described above.
Example four
The present embodiments also provide a computer readable storage medium for storing computer instructions which, when executed by a processor, perform the steps in the method for detecting a thin fracture based on object detection and a graph attention machine as described above.
The steps involved in the second to fourth embodiments correspond to the first embodiment of the method, and the detailed description thereof can be found in the related description of the first embodiment. The term "computer-readable storage medium" should be taken to include a single medium or multiple media containing one or more sets of instructions; it should also be understood to include any medium that is capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any of the methods of the present invention.
It will be understood by those skilled in the art that the modules or steps of the present invention described above may be implemented by a general purpose computer device, and alternatively, they may be implemented by program code executable by a computing device, such that they may be stored in a storage device and executed by the computing device, or they may be separately fabricated into individual integrated circuit modules, or multiple modules or steps thereof may be fabricated into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.
Claims (10)
1. A method for detecting fine fractures based on target detection and a graph attention machine mechanism is characterized by comprising the following steps:
acquiring a plurality of time sequence chest CT images, preprocessing the images, and constructing a training sample set by taking the chest CT images marked with small fracture areas as training samples;
training a tiny fracture detection model by using a training sample set; the training process comprises the following steps: inputting a training sample image into a fine fracture detection model, extracting image characteristics by using a convolutional neural network, extracting a graph attention matrix by using the graph convolutional neural network, combining the image characteristics and the graph attention matrix to obtain final characteristics, outputting a fine fracture region prediction frame by combining a YOLO target detection network based on the final characteristics, and training the fine fracture detection model according to the distance loss between the prediction frame and the target frame;
and inputting the chest CT image to be detected into the trained tiny fracture detection model, and outputting a tiny fracture area and a fracture type.
2. The method for thin fracture detection based on object detection and graph attention mechanism as claimed in claim 1, wherein said preprocessing comprises:
stacking a plurality of time sequence chest CT images to synthesize a multi-channel image;
and resampling according to the inter-pixel distance of the chest CT image to finish normalization.
3. The method of claim 1, wherein the image features and the graph attention matrix are combined to obtain final features, comprising:
calculating a graph attention matrix extracted by a graph convolution neural network by using 1 multiplied by 1 convolution, and changing a final nonlinear activation function into sigmoid to enable a value range of a calculation result to be between 0 and 1;
and multiplying the calculated figure attention force matrix by the characteristic diagram extracted by the convolutional neural network to obtain the final characteristic.
4. The method for detecting fine fractures based on object detection and image attention mechanism according to claim 1, wherein after the time-series chest CT images are inputted into the fine fracture detection model, the corresponding object detection prediction boxes are outputted in sequence, the output result is inputted into the fine fracture detection optimization model, and the overall image sequence prediction result is post-processed and enhanced by using the z-direction-based NMS algorithm and the object tracking algorithm.
5. The method of claim 4, wherein the z-direction based NMS algorithm and the target tracking algorithm are used, and the method comprises:
mapping the position of the prediction frame with the highest confidence level in the previous chest CT image and the position of the prediction frame with the highest confidence level in the next chest CT image to the current chest CT image, taking the position of the prediction frame with the highest confidence level in the current chest CT image as a central point, and comparing the position with the mapping position;
and if the z coordinate of the mapping position is within a preset value range and the IOU of the prediction frame of the mapping position and the prediction frame of the central point position is larger than a set threshold value, deleting the prediction frame of the mapping position, and outputting the prediction frame with the highest confidence level in the current chest CT image as a final prediction frame, otherwise, outputting the prediction frame of the mapping position as the final prediction frame.
6. The method of claim 5, wherein the range of z-coordinate values is determined based on the inter-pixel distance of the CT image of the breast.
7. A tiny fracture detection system based on target detection and a graph attention machine mechanism is characterized by comprising:
the training sample construction module is used for acquiring a plurality of time sequence chest CT images, preprocessing the images, and constructing a training sample set by taking the chest CT images marked with small fracture areas as training samples;
the tiny fracture detection model training module is used for training a tiny fracture detection model by utilizing a training sample set; the training process comprises the following steps: inputting a training sample image into a fine fracture detection model, extracting image characteristics by using a convolutional neural network, extracting a graph attention matrix by using the graph convolutional neural network, combining the image characteristics and the graph attention matrix to obtain final characteristics, outputting a fine fracture region prediction frame by combining a YOLO target detection network based on the final characteristics, and training the fine fracture detection model according to the distance loss between the prediction frame and the target frame;
and the tiny fracture detection module is used for inputting the chest CT image to be detected into a tiny fracture detection model after training and outputting a tiny fracture area and a fracture type.
8. The system according to claim 7, further comprising a minor fracture detection optimization module for inputting the output to the minor fracture detection optimization model, and performing post-processing enhancement on the entire image sequence prediction result by using a z-direction-based NMS algorithm and a target tracking algorithm.
9. An electronic device comprising a memory and a processor and computer instructions stored on the memory and executed on the processor, wherein the computer instructions, when executed by the processor, perform the steps of a method for fine fracture detection based on object detection and graphical attention machine as claimed in any one of claims 1 to 6.
10. A computer readable storage medium storing computer instructions which, when executed by a processor, perform the steps of a method for target-based detection and visualization-based detection of small fractures as claimed in any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211710001.6A CN115908394A (en) | 2022-12-29 | 2022-12-29 | Tiny fracture detection method and system based on target detection and graph attention machine mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211710001.6A CN115908394A (en) | 2022-12-29 | 2022-12-29 | Tiny fracture detection method and system based on target detection and graph attention machine mechanism |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115908394A true CN115908394A (en) | 2023-04-04 |
Family
ID=86479917
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211710001.6A Pending CN115908394A (en) | 2022-12-29 | 2022-12-29 | Tiny fracture detection method and system based on target detection and graph attention machine mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115908394A (en) |
-
2022
- 2022-12-29 CN CN202211710001.6A patent/CN115908394A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107895367B (en) | Bone age identification method and system and electronic equipment | |
| US10706333B2 (en) | Medical image analysis method, medical image analysis system and storage medium | |
| CN110232383B (en) | Focus image recognition method and focus image recognition system based on deep learning model | |
| CN112418329B (en) | Cervical OCT image classification method and system based on multi-scale textural feature fusion | |
| CN106033540B (en) | A kind of microecology in vaginas morphology automatic analysis method and system | |
| WO2022032824A1 (en) | Image segmentation method and apparatus, device, and storage medium | |
| CN113284149B (en) | COVID-19 chest CT image identification method and device and electronic equipment | |
| CN109191465A (en) | A kind of system for being determined based on deep learning network, identifying human body or so the first rib cage | |
| CN109685765A (en) | A kind of X-ray pneumonia prediction of result device based on convolutional neural networks | |
| CN112819831B (en) | Segmentation model generation method and device based on convolution Lstm and multi-model fusion | |
| CN112164026B (en) | Endoscope polyp real-time detection method, system and terminal | |
| CN117078692B (en) | Medical ultrasonic image segmentation method and system based on self-adaptive feature fusion | |
| CN110930373A (en) | Pneumonia recognition device based on neural network | |
| CN110570425B (en) | Pulmonary nodule analysis method and device based on deep reinforcement learning algorithm | |
| CN110895815A (en) | Chest X-ray pneumothorax segmentation method based on deep learning | |
| WO2022086910A1 (en) | Anatomically-informed deep learning on contrast-enhanced cardiac mri | |
| CN116258933A (en) | Medical image segmentation device based on global information perception | |
| CN115359066A (en) | Focus detection method and device for endoscope, electronic device and storage medium | |
| CN115471512A (en) | Medical image segmentation method based on self-supervision contrast learning | |
| CN117036288A (en) | Tumor subtype diagnosis method for full-slice pathological image | |
| CN114565601A (en) | Improved liver CT image segmentation algorithm based on DeepLabV3+ | |
| CN111047582A (en) | Crohn's disease auxiliary diagnosis system under enteroscope based on degree of depth learning | |
| CN112766314A (en) | Anatomical structure recognition method, electronic device, and storage medium | |
| CN115908394A (en) | Tiny fracture detection method and system based on target detection and graph attention machine mechanism | |
| CN115880358A (en) | Construction method of positioning model, positioning method of image mark points and electronic equipment |
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 |