Loop and distillation: Attention weights fusion transformer for fine‐grained representation

Sun, Fayou and Ngo, Hea Choon and Zuqiang, Meng and Sek, Yong Wee (2023) Loop and distillation: Attention weights fusion transformer for fine‐grained representation. IET Computer Vision, 17 (4). pp. 473-482. ISSN 1751-9632

[img] Text
0130221062024102412871.pdf

Download (3MB)

Abstract

Learning subtle discriminative feature representation plays a significant role in Fine-Grained Visual Categorisation (FGVC). The vision transformer (ViT) achieves promising performance in the traditional image classification filed due to its multi-head self-attention mechanism. Unfortunately, ViT cannot effectively capture critical feature regions for FGVC due to only focusing on classification token and adopting the strategy of one-time image input. Besides, the advantage of attention weights fusion is not applied to ViT. To promote the performance of capturing vital regions for FGVC, the authors propose a novel model named RDTrans, which proposes discriminative region with top priority in a recurrent learning way. Specifically, proposed vital regions at each scale will be cropped and amplified as the next input parameters to finally locate the most discriminative region. Furthermore, a distillation learning method is employed to provide better supervision for elevating the generalisation ability. Concurrently, RDTrans can be easily trained end-to-end in a weakly-supervised learning way. Extensive experiments demonstrate that RDTrans yields state-of-the-art performance on four widely used fine-grained benchmarks, including CUB-200-2011, Stanford Cars, Stanford Dogs, and iNat2017.

Item Type: Article
Uncontrolled Keywords: Computer vision, Fine‐grained image recognition, Image processing
Divisions: Faculty of Information and Communication Technology
Depositing User: Norfaradilla Idayu Ab. Ghafar
Date Deposited: 07 Oct 2024 12:41
Last Modified: 07 Oct 2024 12:41
URI: http://eprints.utem.edu.my/id/eprint/27758
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item