Technical Report for Egocentric Mistake Detection for the HoloAssist Challenge

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Patsch, Constantin, Zakour, Marsil, Wu, Yuankai, Steinbach, Eckehard
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910992353460224
author Patsch, Constantin
Zakour, Marsil
Wu, Yuankai
Steinbach, Eckehard
author_facet Patsch, Constantin
Zakour, Marsil
Wu, Yuankai
Steinbach, Eckehard
contents In this report, we address the task of online mistake detection, which is vital in domains like industrial automation and education, where real-time video analysis allows human operators to correct errors as they occur. While previous work focuses on procedural errors involving action order, broader error types must be addressed for real-world use. We introduce an online mistake detection framework that handles both procedural and execution errors (e.g., motor slips or tool misuse). Upon detecting an error, we use a large language model (LLM) to generate explanatory feedback. Experiments on the HoloAssist benchmark confirm the effectiveness of our approach, where our approach is placed second on the mistake detection task.
format Preprint
id arxiv_https___arxiv_org_abs_2506_06174
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Technical Report for Egocentric Mistake Detection for the HoloAssist Challenge
Patsch, Constantin
Zakour, Marsil
Wu, Yuankai
Steinbach, Eckehard
Computer Vision and Pattern Recognition
In this report, we address the task of online mistake detection, which is vital in domains like industrial automation and education, where real-time video analysis allows human operators to correct errors as they occur. While previous work focuses on procedural errors involving action order, broader error types must be addressed for real-world use. We introduce an online mistake detection framework that handles both procedural and execution errors (e.g., motor slips or tool misuse). Upon detecting an error, we use a large language model (LLM) to generate explanatory feedback. Experiments on the HoloAssist benchmark confirm the effectiveness of our approach, where our approach is placed second on the mistake detection task.
title Technical Report for Egocentric Mistake Detection for the HoloAssist Challenge
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.06174