Fine-Grained Zero-Shot Learning with Attribute-Centric Representations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Zhi, Guo, Jingcai, Cai, Taotao, Cai, Yuxiang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912762008961024
author Chen, Zhi
Guo, Jingcai
Cai, Taotao
Cai, Yuxiang
author_facet Chen, Zhi
Guo, Jingcai
Cai, Taotao
Cai, Yuxiang
contents Recognizing unseen fine-grained categories demands a model that can distinguish subtle visual differences. This is typically achieved by transferring visual-attribute relationships from seen classes to unseen classes. The core challenge is attribute entanglement, where conventional models collapse distinct attributes like color, shape, and texture into a single visual embedding. This causes interference that masks these critical distinctions. The post-hoc solutions of previous work are insufficient, as they operate on representations that are already mixed. We propose a zero-shot learning framework that learns AttributeCentric Representations (ACR) to tackle this problem by imposing attribute disentanglement during representation learning. ACR is achieved with two mixture-of-experts components, including Mixture of Patch Experts (MoPE) and Mixture of Attribute Experts (MoAE). First, MoPE is inserted into the transformer using a dual-level routing mechanism to conditionally dispatch image patches to specialized experts. This ensures coherent attribute families are processed by dedicated experts. Finally, the MoAE head projects these expert-refined features into sparse, partaware attribute maps for robust zero-shot classification. On zero-shot learning benchmark datasets CUB, AwA2, and SUN, our ACR achieves consistent state-of-the-art results.
format Preprint
id arxiv_https___arxiv_org_abs_2512_12219
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Fine-Grained Zero-Shot Learning with Attribute-Centric Representations
Chen, Zhi
Guo, Jingcai
Cai, Taotao
Cai, Yuxiang
Computer Vision and Pattern Recognition
Recognizing unseen fine-grained categories demands a model that can distinguish subtle visual differences. This is typically achieved by transferring visual-attribute relationships from seen classes to unseen classes. The core challenge is attribute entanglement, where conventional models collapse distinct attributes like color, shape, and texture into a single visual embedding. This causes interference that masks these critical distinctions. The post-hoc solutions of previous work are insufficient, as they operate on representations that are already mixed. We propose a zero-shot learning framework that learns AttributeCentric Representations (ACR) to tackle this problem by imposing attribute disentanglement during representation learning. ACR is achieved with two mixture-of-experts components, including Mixture of Patch Experts (MoPE) and Mixture of Attribute Experts (MoAE). First, MoPE is inserted into the transformer using a dual-level routing mechanism to conditionally dispatch image patches to specialized experts. This ensures coherent attribute families are processed by dedicated experts. Finally, the MoAE head projects these expert-refined features into sparse, partaware attribute maps for robust zero-shot classification. On zero-shot learning benchmark datasets CUB, AwA2, and SUN, our ACR achieves consistent state-of-the-art results.
title Fine-Grained Zero-Shot Learning with Attribute-Centric Representations
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.12219