Visual detection optimization method and system based on data set classification

By introducing data set classification labels and hierarchical training strategies in visual detection technology, the complexity problem of joint training of multiple data sets in the existing technology is solved, and the adaptability and generalization ability of the model in multiple scenarios is improved.

CN119649130AActive Publication Date: 2025-03-18BEIJING INST OF TECH +1
9 Cites 4 Cited by

Patent Information

Application Number
CN202411781570.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-05
Publication Date
2025-03-18
Estimated Expiration
2044-12-05

AI Technical Summary

Technical Problem

Existing visual detection technology relies on diverse data sets for training and evaluation, with low utilization rate and requires joint training of multiple data sets in complex scenarios, resulting in high training complexity and resource consumption.

Method used

By introducing data set classification labels, hierarchical training strategies and scene adaptive modules are adopted to gradually introduce different scenario data for model training, and hierarchical evaluation and dynamic adjustment of model inference strategies are carried out according to scene classification labels.

Benefits of technology

It realizes good adaptability of the visual detection model in multiple scenarios, simplifies the training process, reduces computing resources and time costs, and enhances the generalization ability and robustness of the model.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a visual detection optimization method and system based on data set classification. The method comprises the steps of introducing a scene classification label and carrying out standardization processing on scene features in a data set labeling and preprocessing stage, and carrying out scene classification multi-task learning in a model selection and training stage, so that a model dynamically adjusts feature learning in different scenes, and the multi-scene adaptive capacity is improved. And in the model evaluation and verification stage, scene-based hierarchical evaluation is introduced and corresponding optimization is carried out. In the model deployment and reasoning stage, a dynamic scene is introduced to adapt to reasoning, and the reasoning path of the model is dynamically adjusted to adapt to the characteristics of the scene, so that the reasoning accuracy is improved. In the later optimization, scene feature clustering and transfer learning are used, and the multi-scene performance of the model is optimized. According to the method, under the condition that the model is not jointly trained through multiple data sets, the multi-scene adaptive capacity of the visual detection model is improved through data set classification.
Need to check novelty before this filing date? Find Prior Art