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

Image training set generation method and device, storage medium and electronic device

A training set and image technology, applied in instruments, computing, character and pattern recognition, etc., can solve the problems of poor low-frequency category recognition effect and scarce image data to be processed, and achieve the effect of improving recognition accuracy and increasing training samples.

Pending Publication Date: 2022-07-29
QINGDAO HAIER TECH +1
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a method, device, storage medium, and electronic device for generating an image training set to at least solve the problem of poor recognition effect of the trained model on low-frequency categories due to the scarcity of image data to be processed in the training samples of low-frequency categories in the related art The problem

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
  • Image training set generation method and device, storage medium and electronic device
  • Image training set generation method and device, storage medium and electronic device
  • Image training set generation method and device, storage medium and electronic device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 2

[0094] According to another embodiment of the present invention, an apparatus for generating an image training set is also provided, image 3 is a block diagram of an image training set generating apparatus according to an embodiment of the present invention, such as image 3 shown, including:

[0095] an acquisition module 32, configured to acquire images of low-frequency categories to be processed from the initial image training set, wherein the low-frequency categories represent that the number of annotation frames of the to-be-processed images satisfies a first preset condition;

[0096] The processing module 34 is used for cropping the to-be-processed image to obtain a target image, and performing image enhancement processing on the target image to obtain a processed target image;

[0097] The replacement module 36 is configured to replace the target area in the image to be processed of the high-frequency category in the initial image training set with the processed targ...

Embodiment 3

[0116] An embodiment of the present invention further provides a storage medium, where a computer program is stored in the storage medium, wherein the computer program is configured to execute the steps in any one of the above method embodiments when running.

[0117] Optionally, in this embodiment, the above-mentioned storage medium may be configured to store a computer program for executing the following steps:

[0118] S1, obtaining images of low-frequency categories to be processed from an initial image training set, wherein the low-frequency categories represent that the number of annotation frames of the to-be-processed images satisfies a first preset condition;

[0119] S2, cropping the to-be-processed image to obtain a target image, and performing image enhancement processing on the target image to obtain a processed target image;

[0120] S3: Replace the target area in the high-frequency category of the image to be processed in the initial image training set with the ...

Embodiment 4

[0123] An embodiment of the present invention also provides an electronic device, comprising a memory and a processor, where a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any of the above method embodiments.

[0124] Optionally, the above-mentioned electronic device may further include a transmission device and an input-output device, wherein the transmission device is connected to the above-mentioned processor, and the input-output device is connected to the above-mentioned processor.

[0125] Optionally, in this embodiment, the above-mentioned processor may be configured to execute the following steps through a computer program:

[0126] S1, obtaining images of low-frequency categories to be processed from an initial image training set, wherein the low-frequency categories represent that the number of annotation frames of the to-be-processed images satisfies a first preset condition;

[0127] S...

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

The invention discloses an image training set generation method and device, a storage medium and an electronic device, and relates to the technical field of smart home / smart home, and the image training set generation method comprises the steps: obtaining a to-be-processed image of a low-frequency category from an initial image training set; cutting the to-be-processed image to obtain a target image, and performing image enhancement processing on the target image to obtain a processed target image; and replacing the target region in the to-be-processed image of the high-frequency category in the initial image training set with the processed target image to obtain a target image training set, thereby solving the problem of poor low-frequency category recognition effect of a trained model caused by rare data of the to-be-processed image of the low-frequency category in a training sample in related technologies. The training samples of the low-frequency category are increased, and the recognition precision of the low-frequency category is improved.

Description

technical field [0001] The present application relates to the technical field of smart home / smart home, and in particular, to a method, device, storage medium and electronic device for generating an image training set. Background technique [0002] In the collection of water wash marks and model training, due to the common long-tail effect of data, the number of categories in the collected training data is always very rare, and there is no way to further increase the number by collecting and crawling, which leads to subsequent models. During training, the recognition of low-frequency (sparse) categories of water-washed labels is very poor. [0003] In the related art, there is no solution to the problem that the image data to be processed of the low-frequency categories in the training samples is sparse, resulting in a poor recognition effect of the trained model for the low-frequency categories. SUMMARY OF THE INVENTION [0004] Embodiments of the present invention 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
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06V10/774G06V10/764G06V10/20G06K9/62
CPCG06F18/24G06F18/214
Inventor 潘威滔
Owner QINGDAO HAIER TECH