DenVisCoM: Dense Vision Correspondence Mamba for Efficient and Real-time Optical Flow and Stereo Estimation

Tushar Anand, Maheswar Bora, Antitza Dantcheva, Abhijit Das · 2026

Read DenVisCoM: Dense Vision Correspondence Mamba for Efficient and Real-time Optical Flow and Stereo Estimation by Tushar Anand, Maheswar Bora, Antitza Dantcheva, Abhijit Das (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01724.txt.

Description

Tushar Anand and colleagues build DenVisCoM, a Mamba block plus Transformer attention hybrid for joint real time optical flow and disparity estimation. Motion and multi view geometry tasks are deeply related, so one unified design tackles them together, with inference speed, memory footprint and accuracy in one. The authors analyze the accuracy versus real time tradeoff across many data sets. Their model estimates optical flow and disparity well in real time, with models and code released.

Catalog record

CREATORTushar Anand, Maheswar Bora, Antitza Dantcheva, Abhijit Das
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE32190
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01724

Browse the full library · About the Tower of Babel