File management system and method for deep learning annotation samples

Abstract

The invention relates to a file management system and method for deep learning annotation samples. The file management system comprises a client and a server in interactive connection with the client.The client comprises a labeled sample renaming module, a labeled sample management module, a labeled sample quality inspection module and a data augmentation module, and the server comprises a database. An original annotation sample is renamed according to a certain naming mode through the annotation sample renaming module, so that a user can select basic information of the annotation sample conveniently and quickly, code analysis data does not need to be developed again, and the analysis and evaluation efficiency of the annotation sample is improved. The annotation sample management module can inquire, download and delete annotation samples in a client file management interface; and the labeled sample quality inspection module is used for inspecting and deleting labeled samples which donot meet the requirements, and the data augmentation module is used for performing data augmentation on the existing labeled samples.The deep learning training sample is rapidly cleaned in effective time, and the working efficiency is improved.

Description

technical field [0001] The present invention relates to the field of computer information management, in particular to a file management system and method for marking samples by deep learning. Background technique [0002] Labeling sample data is the fuel of AI, which fully demonstrates the importance of sample data in the field of automatic driving. Although there are a lot of public data for us to choose from at home and abroad, basically all the sample data are expressed and stored in different ways. These data are all managed manually, and the management method is not uniform enough, which will cause a lot of waste of human resources. , and the speed of manual data entry is slow and the accuracy rate is low. With the continuous development of the scale of autonomous driving, the number and types of samples describing road information are increasing, and the management challenges for sample data are also relatively large. However, the traditional manual operation mode an...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the technical problems existing in the prior art, the present invention provides a file management system and method for deep learning labeling samples, which solves the problem of confusing labeling methods of labeling samples, and realizes the basic information of labeling samples can be quickly understood through the names of labeling samples , improving the efficiency of analysis and evaluation of labeled samples

Images