Object-Centric Cropping for Visual Few-Shot Classification

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Abdali, Aymane, Boguslawski, Bartosz, Drumetz, Lucas, Gripon, Vincent
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912514744254464
author Abdali, Aymane
Boguslawski, Bartosz
Drumetz, Lucas
Gripon, Vincent
author_facet Abdali, Aymane
Boguslawski, Bartosz
Drumetz, Lucas
Gripon, Vincent
contents In the domain of Few-Shot Image Classification, operating with as little as one example per class, the presence of image ambiguities stemming from multiple objects or complex backgrounds can significantly deteriorate performance. Our research demonstrates that incorporating additional information about the local positioning of an object within its image markedly enhances classification across established benchmarks. More importantly, we show that a significant fraction of the improvement can be achieved through the use of the Segment Anything Model, requiring only a pixel of the object of interest to be pointed out, or by employing fully unsupervised foreground object extraction methods.
format Preprint
id arxiv_https___arxiv_org_abs_2508_00218
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Object-Centric Cropping for Visual Few-Shot Classification
Abdali, Aymane
Boguslawski, Bartosz
Drumetz, Lucas
Gripon, Vincent
Computer Vision and Pattern Recognition
Machine Learning
In the domain of Few-Shot Image Classification, operating with as little as one example per class, the presence of image ambiguities stemming from multiple objects or complex backgrounds can significantly deteriorate performance. Our research demonstrates that incorporating additional information about the local positioning of an object within its image markedly enhances classification across established benchmarks. More importantly, we show that a significant fraction of the improvement can be achieved through the use of the Segment Anything Model, requiring only a pixel of the object of interest to be pointed out, or by employing fully unsupervised foreground object extraction methods.
title Object-Centric Cropping for Visual Few-Shot Classification
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2508.00218