Unlock instant, AI-driven research and patent intelligence for your innovation.

A data augmentation method for skin disease images

A skin disease and image technology, applied in the field of computer vision, can solve problems such as affecting the recognition results, and achieve the effect of large interference and good regularization effect.

Active Publication Date: 2021-10-26
ZHEJIANG UNIV OF TECH
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in many image recognition problems, these factors should not affect the final recognition results

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A data augmentation method for skin disease images
  • A data augmentation method for skin disease images

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] The present invention will be further described below in conjunction with the accompanying drawings.

[0030] refer to figure 1 and figure 2 , a data augmentation method for skin disease images, comprising the following steps:

[0031] Step 1: Select a sample m from the skin image sample set M sequentially and randomly without replacement i ;

[0032] Step 2: Randomly sample m from i Select an area with a fixed size for cropping, and the cropped m i named m i1 , from m i The cropped image is named m i2 ;

[0033] Step 3: Randomly select a sample b from the skin image sample set B whose width and height are larger than the width and height of the cropping area with replacement j ;

[0034] Step 4: From sample b j Randomly select a region of fixed size (and step 1 from m i The size of the above cropped area is the same) to crop, and the cropped b j named b j1 , the sample cropped from bj is named b j2 ;

[0035]Step 5: according to formula c=λ*value (m i...

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

No PUM Login to View More

Abstract

A data enhancement method for skin disease images, comprising the following steps: Step 1: Select a sample m from the skin image sample set M sequentially and randomly without replacement i ; Step 2: Randomly start from sample m i Select an area with a fixed size for cropping, and the cropped m i named m i1 , from m i The cropped image is named m i2 ; Step 3: Randomly select a sample b from the skin image sample set B whose width and height are larger than the width and height of the clipping area with replacement j ; Step 4: From sample b j Randomly select a fixed-size area for cropping, and the cropped b j named b j1 , the sample cropped from bj is named b j2 ; Step 5: Calculate c, round the non-integer pixel value in the result; Step 6: c and m i1 Fusion is performed to form a new m i . The invention provides a data enhancement method for skin disease images to improve the performance of a skin disease classification model.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a novel method for data enhancement of skin disease images. Specifically, the size of the data set is expanded by cutting out different types of skin disease images and then merging them to generate new samples. Background technique [0002] Data enhancement is a commonly used data preprocessing method in deep learning. Through simple data enhancement, we can prevent the model from overfitting and improve the generalization of the model. In multiple deep learning frameworks such as tensorflow, caffe, pytorch, etc., some simple and commonly used data enhancements are integrated. Through these data enhancements, we can quickly expand the data set and expand the data set to a certain level. For example, use mirror in caffe to mirror the original image, use crop to crop the image according to the set size, tensorflow also integrates more data enhancement methods, and python...

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
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/00G06T5/00G06T5/50
CPCG06T5/00G06T5/50G06T7/0012G06T2207/20221G06T2207/30088G06T2210/22
Inventor 胡海根孔祥勇苏一平陈胜勇周乾伟管秋
Owner ZHEJIANG UNIV OF TECH