Sample data cleaning method, device, computer device, and storage medium

Abstract

The invention discloses a sample data cleaning method, a device, a computer device and a storage medium. Firstly, an initial image set is obtained, and the initial image set is input to a feature classification model for recognition to obtain a feature recognition result. Then, according to the feature recognition result, an initial training set is obtained, and the initial training set comprisesan initial training image and labeling data corresponding to each initial training image. In order to ensure the data richness of the training data, the initial training set is obtained by selecting the image data in the initial image set according to the preset requirements through the feature classification model. On this basis, the initial training images are classified according to the annotation data, and the classification training set is obtained. And the classification training set is cleaned to obtain the target training set. On the basis of ensuring the data richness of the trainingdata, the training data is cleaned to ensure the accuracy of the training data, in order to further improve the accuracy of the follow-up model training.

Description

technical field [0001] The present invention relates to the field of data processing, in particular to a sample data cleaning method, device, computer equipment and storage medium. Background technique [0002] With the development of computer technology, deep learning has been widely used in various fields. The training of deep learning requires a large number of training samples. If there are fewer training samples, the effect will be much worse. For training samples, traditionally, the existing training sample data is obtained from the network or a third-party data platform for model training. However, for many specific application scenarios, the existing training sample data in these networks or third-party data platforms may not be able to meet actual needs, so many sample data need to be collected manually, which makes it very inconvenient to obtain training sample data . Moreover, due to the relatively large amount of data in the training samples, there are often e...

AI Technical Summary

Problems solved by technology

[0003] Embodiments of the present invention provide a sample data cleaning method, device, computer equipment, and storage medium to solve the problem of low accuracy of training sample data

Images