DispBench : Benchmarking Disparity Estimation to Synthetic Corruptions

Shashank Agnihotri, Amaan Ansari, Annika Dackermann, Fabian Rösch, Margret Keuper · 2025

Read DispBench : Benchmarking Disparity Estimation to Synthetic Corruptions by Shashank Agnihotri, Amaan Ansari, Annika Dackermann, Fabian Rösch, Margret Keuper (2025) 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-2505-05091.txt.

Description

Shashank Agnihotri and four co-authors introduce DispBench in a 2025 paper, a benchmark for disparity estimation under synthetic corruptions. Disparity estimation finds the shift between matched pixels in a stereo pair of images. It feeds safety-critical work such as medical surgery and self-driving.

Deep learning tools beat humans on clean test sets, yet they can crack when data shift or under attack. No standard test of their robustness was on hand. DispBench fills that gap: it scores tools in a set way across made-up flaws, so groups can see how well a method holds up when the world is less than clean.

Catalog record

CREATORShashank Agnihotri, Amaan Ansari, Annika Dackermann, Fabian Rösch, Margret Keuper
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-05091

Browse the full library · About the Tower of Babel