ClearCLIP: Decomposing CLIP Representations for Dense Vision-Language Inference

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Lan, Mengcheng, Chen, Chaofeng, Ke, Yiping, Wang, Xinjiang, Feng, Litong, Zhang, Wayne
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916328284094464
author Lan, Mengcheng
Chen, Chaofeng
Ke, Yiping
Wang, Xinjiang
Feng, Litong
Zhang, Wayne
author_facet Lan, Mengcheng
Chen, Chaofeng
Ke, Yiping
Wang, Xinjiang
Feng, Litong
Zhang, Wayne
contents Despite the success of large-scale pretrained Vision-Language Models (VLMs) especially CLIP in various open-vocabulary tasks, their application to semantic segmentation remains challenging, producing noisy segmentation maps with mis-segmented regions. In this paper, we carefully re-investigate the architecture of CLIP, and identify residual connections as the primary source of noise that degrades segmentation quality. With a comparative analysis of statistical properties in the residual connection and the attention output across different pretrained models, we discover that CLIP's image-text contrastive training paradigm emphasizes global features at the expense of local discriminability, leading to noisy segmentation results. In response, we propose ClearCLIP, a novel approach that decomposes CLIP's representations to enhance open-vocabulary semantic segmentation. We introduce three simple modifications to the final layer: removing the residual connection, implementing the self-self attention, and discarding the feed-forward network. ClearCLIP consistently generates clearer and more accurate segmentation maps and outperforms existing approaches across multiple benchmarks, affirming the significance of our discoveries.
format Preprint
id arxiv_https___arxiv_org_abs_2407_12442
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ClearCLIP: Decomposing CLIP Representations for Dense Vision-Language Inference
Lan, Mengcheng
Chen, Chaofeng
Ke, Yiping
Wang, Xinjiang
Feng, Litong
Zhang, Wayne
Computer Vision and Pattern Recognition
Despite the success of large-scale pretrained Vision-Language Models (VLMs) especially CLIP in various open-vocabulary tasks, their application to semantic segmentation remains challenging, producing noisy segmentation maps with mis-segmented regions. In this paper, we carefully re-investigate the architecture of CLIP, and identify residual connections as the primary source of noise that degrades segmentation quality. With a comparative analysis of statistical properties in the residual connection and the attention output across different pretrained models, we discover that CLIP's image-text contrastive training paradigm emphasizes global features at the expense of local discriminability, leading to noisy segmentation results. In response, we propose ClearCLIP, a novel approach that decomposes CLIP's representations to enhance open-vocabulary semantic segmentation. We introduce three simple modifications to the final layer: removing the residual connection, implementing the self-self attention, and discarding the feed-forward network. ClearCLIP consistently generates clearer and more accurate segmentation maps and outperforms existing approaches across multiple benchmarks, affirming the significance of our discoveries.
title ClearCLIP: Decomposing CLIP Representations for Dense Vision-Language Inference
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.12442