Saved in:
Bibliographic Details
Main Authors: Mededovic, Emil, Wu, Yuli, Konermann, Henning, Kopaczka, Marcin, Schulz, Mareike, Tolba, Rene, Stegmaier, Johannes
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.10305
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909536695091200
author Mededovic, Emil
Wu, Yuli
Konermann, Henning
Kopaczka, Marcin
Schulz, Mareike
Tolba, Rene
Stegmaier, Johannes
author_facet Mededovic, Emil
Wu, Yuli
Konermann, Henning
Kopaczka, Marcin
Schulz, Mareike
Tolba, Rene
Stegmaier, Johannes
contents Analyzing animal behavior from video recordings is crucial for scientific research, yet manual annotation remains labor-intensive and prone to subjectivity. Efficient segmentation methods are needed to automate this process while maintaining high accuracy. In this work, we propose a novel pipeline that utilizes eye-tracking data from Aria glasses to generate prompt points, which are then used to produce segmentation masks via a fast zero-shot segmentation model. Additionally, we apply post-processing to refine the prompts, leading to improved segmentation quality. Through our approach, we demonstrate that combining eye-tracking-based annotation with smart prompt refinement can enhance segmentation accuracy, achieving an improvement of 70.6% from 38.8 to 66.2 in the Jaccard Index for segmentation results in the rats dataset.
format Preprint
id arxiv_https___arxiv_org_abs_2503_10305
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Eye on the Target: Eye Tracking Meets Rodent Tracking
Mededovic, Emil
Wu, Yuli
Konermann, Henning
Kopaczka, Marcin
Schulz, Mareike
Tolba, Rene
Stegmaier, Johannes
Computer Vision and Pattern Recognition
Analyzing animal behavior from video recordings is crucial for scientific research, yet manual annotation remains labor-intensive and prone to subjectivity. Efficient segmentation methods are needed to automate this process while maintaining high accuracy. In this work, we propose a novel pipeline that utilizes eye-tracking data from Aria glasses to generate prompt points, which are then used to produce segmentation masks via a fast zero-shot segmentation model. Additionally, we apply post-processing to refine the prompts, leading to improved segmentation quality. Through our approach, we demonstrate that combining eye-tracking-based annotation with smart prompt refinement can enhance segmentation accuracy, achieving an improvement of 70.6% from 38.8 to 66.2 in the Jaccard Index for segmentation results in the rats dataset.
title Eye on the Target: Eye Tracking Meets Rodent Tracking
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.10305