Learning to Adapt Category Consistent Meta-Feature of CLIP for Few-Shot Classification
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913420771590144 |
|---|---|
| author | Shi, Jiaying Xue, Xuetong Xu, Shenghui |
| author_facet | Shi, Jiaying Xue, Xuetong Xu, Shenghui |
| contents | The recent CLIP-based methods have shown promising zero-shot and few-shot performance on image classification tasks. Existing approaches such as CoOp and Tip-Adapter only focus on high-level visual features that are fully aligned with textual features representing the ``Summary" of the image. However, the goal of few-shot learning is to classify unseen images of the same category with few labeled samples. Especially, in contrast to high-level representations, local representations (LRs) at low-level are more consistent between seen and unseen samples. Based on this point, we propose the Meta-Feature Adaption method (MF-Adapter) that combines the complementary strengths of both LRs and high-level semantic representations. Specifically, we introduce the Meta-Feature Unit (MF-Unit), which is a simple yet effective local similarity metric to measure category-consistent local context in an inductive manner. Then we train an MF-Adapter to map image features to MF-Unit for adequately generalizing the intra-class knowledge between unseen images and the support set. Extensive experiments show that our proposed method is superior to the state-of-the-art CLIP downstream few-shot classification methods, even showing stronger performance on a set of challenging visual classification tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_05647 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Learning to Adapt Category Consistent Meta-Feature of CLIP for Few-Shot Classification Shi, Jiaying Xue, Xuetong Xu, Shenghui Computer Vision and Pattern Recognition The recent CLIP-based methods have shown promising zero-shot and few-shot performance on image classification tasks. Existing approaches such as CoOp and Tip-Adapter only focus on high-level visual features that are fully aligned with textual features representing the ``Summary" of the image. However, the goal of few-shot learning is to classify unseen images of the same category with few labeled samples. Especially, in contrast to high-level representations, local representations (LRs) at low-level are more consistent between seen and unseen samples. Based on this point, we propose the Meta-Feature Adaption method (MF-Adapter) that combines the complementary strengths of both LRs and high-level semantic representations. Specifically, we introduce the Meta-Feature Unit (MF-Unit), which is a simple yet effective local similarity metric to measure category-consistent local context in an inductive manner. Then we train an MF-Adapter to map image features to MF-Unit for adequately generalizing the intra-class knowledge between unseen images and the support set. Extensive experiments show that our proposed method is superior to the state-of-the-art CLIP downstream few-shot classification methods, even showing stronger performance on a set of challenging visual classification tasks. |
| title | Learning to Adapt Category Consistent Meta-Feature of CLIP for Few-Shot Classification |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2407.05647 |