DCNN: Dual Cross-current Neural Networks Realized Using An Interactive Deep Learning Discriminator for Fine-grained Objects

TL;DR

This paper introduces DCNN, a dual neural network combining convolution and self-attention to enhance fine-grained image classification accuracy, outperforming existing methods on benchmark datasets.

Contribution

The study presents a novel dual-current neural network architecture that fuses convolutional and self-attention features for improved fine-grained image classification.

Findings

Achieved 13.5-19.5% performance improvements on benchmark datasets.

Demonstrated effectiveness of combining convolution and self-attention mechanisms.

Proposed a weakly supervised backbone model with enhanced feature extraction.

Abstract

Accurate classification of fine-grained images remains a challenge in backbones based on convolutional operations or self-attention mechanisms. This study proposes novel dual-current neural networks (DCNN), which combine the advantages of convolutional operations and self-attention mechanisms to improve the accuracy of fine-grained image classification. The main novel design features for constructing a weakly supervised learning backbone model DCNN include (a) extracting heterogeneous data, (b) keeping the feature map resolution unchanged, (c) expanding the receptive field, and (d) fusing global representations and local features. Experimental results demonstrated that using DCNN as the backbone network for classifying certain fine-grained benchmark datasets achieved performance advantage improvements of 13.5--19.5% and 2.2--12.9%, respectively, compared to other advanced convolution or attention-based fine-grained backbones.