An image recognition model training and image recognition method, device and system
An image recognition and recognition model technology, applied in the computer field, can solve the problems of high cost and long time, and achieve the effect of improving efficiency, reducing cost and improving accuracy
Active Publication Date: 2020-11-03
TENCENT TECH (SHENZHEN) CO LTD
View PDF9 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0002] At present, the training methods used in the endoscopic imaging diagnosis system use pre-labeled data in the iterative process, requiring a large amount of labeled data, and these labeled data need to be labeled by doctors or experts, and the cost is relatively high. High and time-consuming, there is no relevant solution for this
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment approach
[0096] The first implementation manner: according to each marked video image and the initial training image sample set, the image recognition model is obtained through retraining.
[0097] That is to say, in the embodiment of the present application, all previous initial training image sample sets and labeled image data can be combined to retrain the image recognition model. Since this method needs to restart the training of the image recognition model, the time cost may be relatively low. High, but it can ensure that the model training process fits the full amount of data.
[0098] The second implementation manner: update and train the initial image recognition model according to each labeled video image and the initial training image sample set.
[0099] That is to say, in the embodiment of the present application, iterative update can also be performed on the basis of the initial image recognition model, and the initial image recognition model can be further fine-tuned. For...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The present application relates to the field of computer technology, and in particular to an image recognition model training and image recognition method, device and system. Each image is recognized according to the initial image recognition model, and the predicted lesion category of each image is obtained respectively, and according to each image The image report associated with the image judges the predicted lesion category, labels the lesion category of each image image according to the judgment result, and obtains the image recognition model through iterative training based on each labeled image image and the initial training image sample set, and then can be based on the training. The image recognition model recognizes the lesion category of the image to be recognized, and can determine the lesion category recognition result of the image to be recognized. In this way, using the image report for iterative training does not require additional labeling costs, improves the iteration rate, and can be updated with continuous iteration , to improve the recognition accuracy.
Description
technical field [0001] The present application relates to the field of computer technology, in particular to an image recognition model training and image recognition method, device and system. Background technique [0002] At present, the training methods used in the endoscopic imaging diagnosis system use pre-labeled data in the iterative process, requiring a large amount of labeled data, and these labeled data need to be labeled by doctors or experts, and the cost is relatively high. High and time-consuming, there is no relevant solution for this. Contents of the invention [0003] Embodiments of the present application provide an image recognition model training and image recognition method, device and system, so as to increase the iteration rate of model training for lesion recognition and reduce costs. [0004] The specific technical scheme that the embodiment of the present application provides is as follows: [0005] An embodiment of the present application provi...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/62
CPCG06V2201/03G06F18/24G06F18/214
Inventor 郑瀚尚鸿孙钟前
Owner TENCENT TECH (SHENZHEN) CO LTD