Data enhancement method applied to medicine recognition
A data and medicine technology, applied in the field of data enhancement, can solve the problem of ineffective model improvement, avoid over-fitting problems, improve generalization ability, and achieve the effect of small correlation
Active Publication Date: 2019-06-21
颐保医疗科技(上海)有限公司
View PDF5 Cites 3 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
These enhanced image samples will have a good effect in the initial stage, but because the new image generated by the traditional data enhancement technology is only slightly changed on the original image, the generated image and the original image have some local differences. The information contained is still completely consistent. As the data enhancement continues, the effect of increasing the number of samples on improving the model will become less and less obvious
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 Construction
[0022] The present invention will be further described below in combination with specific embodiments and accompanying drawings.
[0023] like Figure 1 to Figure 5 As shown, a data enhancement method applied to medical identification is disclosed, which is a method different from traditional data enhancement techniques. First, use the GAN to enhance the data of the image. Unlike the traditional data enhancement, the new image generated by the GAN does not do some linear transformation on the original image, but uses the original image sample as the training sample of the GAN. The characteristics of each category of the original image are used to generate some new images that do not exist and satisfy the original image classification. Also, replacing the background color of an image is a good data augmentation method. Generally speaking, image pixel values are expressed in RGB format. The HSV format is a common model for expressing color in cylindrical coordinates, which ...
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 invention discloses a data enhancement method applied to medicine identification, which comprises the following steps of: 1, performing data enhancement on pictures in a training set by using a traditional data enhancement technology to obtain a new picture; 2, constructing a new picture generated by the generative adversarial network; 3, carrying out data enhancement on the training set picture by using the replacement background color, and storing a new picture; and 4, combining the pictures of the original training set with the new pictures generated by the traditional data enhancementtechnology, the new pictures generated by the generative adversarial network and the new pictures generated by the background color replacement technology to form a training sample, and sending the training sample into the traditional Chinese medicinal material recognition artificial intelligence model for training. The correlation between the generated new picture and the original picture is small, by using the data enhancement technology, the purpose of expanding the data set can be achieved, meanwhile, the overfitting problem possibly caused by a traditional data enhancement technology is avoided, and the artificial intelligence model based on the traditional Chinese medicine picture data set has better generalization capability.
Description
technical field [0001] The invention relates to a data enhancement method, in particular to a data enhancement method applied to medical identification. Background technique [0002] In recent years, with the rapid development of artificial intelligence and big data technology, the application of artificial intelligence algorithms in the field of identification of traditional Chinese medicinal materials is also becoming more and more popular. Existing artificial intelligence technologies, such as image recognition technology, use deep learning algorithms. The deep neural network has many parameters and complex structure, which requires a large number of samples for model training. Common things such as cars and faces can obtain a large number of image samples from the Internet, but due to the particularity of Chinese medicinal materials, all image samples need to be taken by themselves, so the number of samples is usually insufficient. [0003] Data augmentation is a commo...
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 Applications(China)
IPC IPC(8): G06T5/50
Inventor 袁杨许慧张群华
Owner 颐保医疗科技(上海)有限公司