Weak supervision fine-grained image recognition method based on visual self-attention mechanism

Abstract

The invention discloses a weak supervision fine-grained image recognition method based on a visual self-attention mechanism. The method involves a student-model module, a teacher-model module and a classification-model module. The student-model and the teacher-model are combined through a Teacher-Student loop feedback mechanism based on Pairwise Approach sorting learning so as to form a self-attention region recommendation network, so that the relation between discriminative region positioning and fine-grained feature learning is enhanced, the discriminative region in the fine-grained image can still be accurately detected under the condition of lacking a target bounding box and a part marking point, and the recognition accuracy is promoted to be remarkably improved; meanwhile, a convolution layer is shared by the three modules, namely, the student-model, the teacher-model and the class-model, so that the model storage space is effectively compressed, the calculation cost is reduced, the method meets the real-time recognition task requirement, and the method is suitable for a large-scale real scene; and besides, a dynamic weight distribution mechanism is adopted in multi-task jointlearning to reduce the amount of artificially set hyper-parameters and enhance the robustness of the model, finally, the whole model is trained and learned in an end-to-end single-stage mode, and thenetwork optimization difficulty is reduced.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a weakly supervised fine-grained image recognition method based on a visual self-attention mechanism. Background technique [0002] Fine-grained image recognition is a challenging research topic in the field of computer vision, which aims to distinguish different sub-categories of the same general category. Compared with the cross-species coarse-grained image recognition task, the appearance similarity of different kinds of objects in the fine-grained image dataset is higher, and the same kind of objects have significant visual differences due to factors such as illumination, pose, and perspective, so the general depth is used. Learning image recognition techniques has difficulty in accurately discriminating fine-grained object categories. Previous studies have shown that the difficulties of fine-grained image recognition tasks lie in discriminative region localization a...

AI Technical Summary

Problems solved by technology

This training method is more complicated, and it is easy to fall into a local optimal solution.

[0005] In summary, for weakly supervised fine-grained image recognition tasks that only use image-level label information, existing methods are difficult to detect image discriminative regions and learn fine-grained features simply, accurately and stably, so an adaptive A fine-grained image recognition method based on visual self-attention mechanism with high robustness

Images