Frozen CLIP: A Strong Backbone for Weakly Supervised Semantic Segmentation

TL;DR

This paper introduces WeCLIP, a novel weakly supervised semantic segmentation method that leverages a frozen CLIP model as the backbone, combined with a new decoder and refinement module, achieving superior performance with reduced training cost.

Contribution

WeCLIP is the first to directly use a frozen CLIP model as the backbone for weakly supervised segmentation, integrating a new decoder and refinement module for improved accuracy.

Findings

Outperforms existing weakly supervised methods significantly.

Reduces training cost compared to prior approaches.

Achieves promising results even in fully supervised settings.

Abstract

Weakly supervised semantic segmentation has witnessed great achievements with image-level labels. Several recent approaches use the CLIP model to generate pseudo labels for training an individual segmentation model, while there is no attempt to apply the CLIP model as the backbone to directly segment objects with image-level labels. In this paper, we propose WeCLIP, a CLIP-based single-stage pipeline, for weakly supervised semantic segmentation. Specifically, the frozen CLIP model is applied as the backbone for semantic feature extraction, and a new decoder is designed to interpret extracted semantic features for final prediction. Meanwhile, we utilize the above frozen backbone to generate pseudo labels for training the decoder. Such labels cannot be optimized during training. We then propose a refinement module (RFM) to rectify them dynamically. Our architecture enforces the proposed decoder and RFM to benefit from each other to boost the final performance. Extensive experiments show that our approach significantly outperforms other approaches with less training cost. Additionally, our WeCLIP also obtains promising results for fully supervised settings. The code is available at https://github.com/zbf1991/WeCLIP.