SemanticMIM: Marring Masked Image Modeling with Semantics Compression for General Visual Representation

TL;DR

SemanticMIM is a novel framework that combines masked image modeling and contrastive learning to improve general visual representations by leveraging their complementary strengths in compression and reconstruction.

Contribution

It introduces a proxy architecture that effectively integrates MIM and CL, enhancing semantic awareness and interpretability in visual representations.

Findings

Significant performance improvements over existing methods.

Enhanced feature linear separability and semantic understanding.

Improved interpretability through attention visualization.

Abstract

This paper represents a neat yet effective framework, named SemanticMIM, to integrate the advantages of masked image modeling (MIM) and contrastive learning (CL) for general visual representation. We conduct a thorough comparative analysis between CL and MIM, revealing that their complementary advantages fundamentally stem from two distinct phases, i.e., compression and reconstruction. Specifically, SemanticMIM leverages a proxy architecture that customizes interaction between image and mask tokens, bridging these two phases to achieve general visual representation with the property of abundant semantic and positional awareness. Through extensive qualitative and quantitative evaluations, we demonstrate that SemanticMIM effectively amalgamates the benefits of CL and MIM, leading to significant enhancement of performance and feature linear separability. SemanticMIM also offers notable interpretability through attention response visualization. Codes are available at https://github.com/yyk-wew/SemanticMIM.