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

Training method, device and equipment and storage medium

A training method and a technology of training samples, which are applied in the computer field, can solve the problems of reducing the image classification performance of an image classifier, reducing the classification accuracy of an image classifier, and incorrectly identifying images, etc.

Active Publication Date: 2020-12-25
北京远鉴信息技术有限公司
View PDF5 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] For the semi-supervised training method, since the training data used by the image classifier may include wrong training data, the trained image classifier still cannot correctly identify images that cannot be correctly identified, and it will also reduce the performance of the image classifier. The classification performance of this type of image reduces the classification accuracy of the image classifier

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
  • Training method, device and equipment and storage medium
  • Training method, device and equipment and storage medium
  • Training method, device and equipment and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0068] figure 1 A schematic flow chart of a training method provided in Embodiment 1 of the present application, such as figure 1 As shown, the method includes the following steps:

[0069] Step 101. Use a classifier to be trained to classify the first candidate sample set to obtain a first classification result of each image included in the first candidate sample set, wherein the number of classification categories of the classifier to be trained is K, K is a positive integer.

[0070] Specifically, collect a certain amount of image data in advance, for example: collect a certain amount of image data on the network, and then use a convolutional neural network to train an image classifier. If you want to obtain an image classifier that can recognize K categories, then Select K categories of images from the collected image data to train the convolutional neural network. The number of images in each category can be 1000-2000. For example: want to get an image classification th...

Embodiment 2

[0101] Figure 5 A schematic structural diagram of a training device provided in Embodiment 2 of the present application, such as Figure 5 As shown, the device includes:

[0102] The first classification unit 51 is configured to use a classifier to be trained to classify the first candidate sample set to obtain a first classification result of each image included in the first candidate sample set, wherein the classification of the classifier to be trained The number of categories is K, and K is a positive integer;

[0103] The first determining unit 52 is configured to determine each image included in the first candidate sample set according to the first classification result and the obtained second manual classification result of each image included in the first candidate sample set sample type, wherein the sample type includes a first sample and a second sample, the first sample is an image with the same first classification result and the second classification result, an...

Embodiment 3

[0125] Figure 8 A schematic structural diagram of an electronic device provided in Embodiment 3 of the present application, including: a processor 801, a storage medium 802, and a bus 803. The storage medium 802 stores machine-readable instructions executable by the processor 801. When When the electronic device runs the data processing method of the consortium chain mentioned above, the processor 801 communicates with the storage medium 802 through the bus 803, and the processor 801 executes the machine-readable instructions to execute the data in the first embodiment. The method steps described.

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 provides a training method and device, equipment and a storage medium, and the method comprises the steps: carrying out the classification of a first candidate sample set through a to-be-trained classifier, and obtaining a first classification result of each image in the first candidate sample set; determining the sample type of each image included in the first candidate sample set according to the first classification result and the obtained second classification result of each image included in the first candidate sample set; selecting an image from the first sample and the second sample according to a preset first sample type proportion, and taking the selected image as a first training sample; and performing model training on the to-be-trained classifier by using the first training sample. Through the method, the classification performance of the to-be-trained classifier can be improved, and the classification accuracy of the to-be-trained classifier can be improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular, to a training method, device, equipment and storage medium. Background technique [0002] Image classifiers can classify images, and image classifiers have important applications in the image field. The classification performance of the image classifier is closely related to the training data. When the image classifier is trained, a large amount of training data needs to be used, and before the image classifier is trained, the training data needs to be manually labeled, and the manual label This method greatly reduces the training efficiency. [0003] In order to improve the training efficiency, the image classifier is currently trained in a semi-supervised manner, for example: after obtaining the training data, use a manual method to label a small part of the training data, and then use the labeled training data to classify the image After the training is completed, ...

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): G06K9/62
CPCG06F18/22G06F18/23G06F18/214G06F18/24
Inventor 王学彬
Owner 北京远鉴信息技术有限公司