Fine-grained image classification model training method and system based on heterogeneous resolution

By constructing a fine-grained image classification model training method with heterogeneous resolutions, multi-scale image samples are generated using bilinear downsampling and mosaic processing. By combining normalized flow networks and adaptive flow networks, the problem of feature loss in mixed-resolution image training is solved, achieving efficient feature repair and alignment, and improving the robustness and generalization ability of the model.

CN121884005BActive Publication Date: 2026-05-15SHANDONG UNIV
2 Cites 0 Cited by

Patent Information

Application Number
CN202610327458.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-03-18
Publication Date
2026-05-15
Estimated Expiration
2046-03-18

AI Technical Summary

Technical Problem

Existing fine-grained image classification models struggle to learn effective features when faced with images of mixed resolutions, resulting in insufficient robustness and generalization ability in complex scenarios, thus failing to meet the requirements for high-precision classification.

Method used

By constructing a fine-grained image classification model training method with heterogeneous resolution, multi-scale image samples are generated using bilinear downsampling and mosaic processing. By combining normalized flow networks and adaptive flow networks, feature repair and alignment are achieved through information density maps and semantic manifold space constraints, thereby optimizing model parameters.

Benefits of technology

Without increasing computational costs, the model's robustness and generalization ability to multi-resolution inputs are improved, ensuring high-precision classification performance in complex environments.

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

Abstract

The application provides a fine-grained image classification model training method and system based on heterogeneous resolution, and relates to the technical field of model training. Based on high-resolution fine-grained image samples to multi-scale images; mosaic processing is performed on the multi-scale images to obtain simulated degradation heterogeneous image samples; the images are input into a fine-grained image classification model, wherein the high-resolution fine-grained image samples are input into a canonical stream network to obtain canonical features, and the heterogeneous image samples are input into an adaptive stream network to obtain adaptive features; an information density map is generated based on the energy difference of the canonical features and the adaptive features, the information density map is used as a weight to construct a feature repair loss, and the adaptive features are guided to align with the canonical features; the adaptive features and the canonical features are respectively mapped to a semantic manifold space, and the consistency of the adaptive features and the canonical features in the semantic direction is constrained; model parameters are optimized by a joint loss function, and a trained fine-grained image classification model is obtained, so that the robustness of the model and the accuracy of the model applied to image classification are improved.
Need to check novelty before this filing date? Find Prior Art