Saved in:
Bibliographic Details
Main Authors: Tore, Onur Berk, Yalciner, Ibrahim Samil, Calap, Server
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.20335
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911286223175680
author Tore, Onur Berk
Yalciner, Ibrahim Samil
Calap, Server
author_facet Tore, Onur Berk
Yalciner, Ibrahim Samil
Calap, Server
contents Estimating homography from a single image remains a challenging yet practically valuable task, particularly in domains like retail, where only one viewpoint is typically available for shelf monitoring and product alignment. In this paper, we present a deep learning framework that predicts a 4-point parameterized homography matrix to rectify shelf images captured from arbitrary angles. Our model leverages a ConvNeXt-based backbone for enhanced feature representation and adopts normalized coordinate regression for improved stability. To address data scarcity and promote generalization, we introduce a novel augmentation strategy by modeling and sampling synthetic homographies. Our method achieves a mean corner error of 1.298 pixels on the test set. When compared with both classical computer vision and deep learning-based approaches, our method demonstrates competitive performance in both accuracy and inference speed. Together, these results establish our approach as a robust and efficient solution for realworld single-view rectification. To encourage further research in this domain, we will make our dataset, ShelfRectSet, and code publicly available
format Preprint
id arxiv_https___arxiv_org_abs_2511_20335
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ShelfRectNet: Single View Shelf Image Rectification with Homography Estimation
Tore, Onur Berk
Yalciner, Ibrahim Samil
Calap, Server
Computer Vision and Pattern Recognition
Estimating homography from a single image remains a challenging yet practically valuable task, particularly in domains like retail, where only one viewpoint is typically available for shelf monitoring and product alignment. In this paper, we present a deep learning framework that predicts a 4-point parameterized homography matrix to rectify shelf images captured from arbitrary angles. Our model leverages a ConvNeXt-based backbone for enhanced feature representation and adopts normalized coordinate regression for improved stability. To address data scarcity and promote generalization, we introduce a novel augmentation strategy by modeling and sampling synthetic homographies. Our method achieves a mean corner error of 1.298 pixels on the test set. When compared with both classical computer vision and deep learning-based approaches, our method demonstrates competitive performance in both accuracy and inference speed. Together, these results establish our approach as a robust and efficient solution for realworld single-view rectification. To encourage further research in this domain, we will make our dataset, ShelfRectSet, and code publicly available
title ShelfRectNet: Single View Shelf Image Rectification with Homography Estimation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.20335