ViM-Disparity: Bridging the Gap of Speed, Accuracy and Memory for Disparity Map Generation

TL;DR

This paper introduces ViM-Disparity, a novel architecture that balances speed, accuracy, and memory efficiency for disparity map generation, along with a new performance measure to evaluate these aspects jointly.

Contribution

The work presents a Visual Mamba-based architecture that overcomes the trade-offs in real-time disparity map generation and proposes a comprehensive performance measure.

Findings

Achieves real-time disparity map generation with high accuracy

Reduces computation overhead compared to existing methods

Provides a new joint evaluation metric for speed, accuracy, and memory

Abstract

In this work we propose a Visual Mamba (ViM) based architecture, to dissolve the existing trade-off for real-time and accurate model with low computation overhead for disparity map generation (DMG). Moreover, we proposed a performance measure that can jointly evaluate the inference speed, computation overhead and the accurateness of a DMG model. The code implementation and corresponding models are available at: https://github.com/MBora/ViM-Disparity.