AutoDOM: Automated Dimension Overlay for Enhanced Measurement-Guidance


Pushpendu Ghosh (Amazon), Aniket Joshi (Amazon), Soumyajit Chowdhury (Amazon), Promod Yenigalla (Amazon)
The 35th British Machine Vision Conference
PDFPosterVideo (Right click to download)

Abstract

Customers looking for furniture products online find it challenging to understand product dimensions, hindering their ability to envision the fit within their spaces. Addressing this, we present a novel automated approach to overlay dimensional lines onto product images, empowering users to understand each subcomponent’s size and scale. Our proposed multi-stage approach uses 3 key components: 3DBoundDetector to identify a bounding box around the product, QuadDetector to identify product subcomponents and AlignMatic, a post processing algorithm to determine optimal placement of dimensional lines to be overlaid. Additionally, we devise an AutoQA mechanism which ensures high-quality and accurate dimension lines by filtering aesthetically poor and incorrect dimension overlays, achieving 91.14% acceptability rate at 50% actionability, thus significantly elevating the customer experience. We benchmark our methodology against state-of-the-art image generation techniques and present ablation studies emphasizing each component's importance within our pipeline.

Citation

@inproceedings{Ghosh_2024_BMVC,
author    = {Pushpendu Ghosh and Aniket Joshi and Soumyajit Chowdhury and Promod Yenigalla},
title     = {AutoDOM: Automated Dimension Overlay for Enhanced Measurement-Guidance},
booktitle = {35th British Machine Vision Conference 2024, {BMVC} 2024, Glasgow, UK, November 25-28, 2024},
publisher = {BMVA},
year      = {2024},
url       = {https://papers.bmvc2024.org/0731.pdf}
}


Copyright © 2024 The British Machine Vision Association and Society for Pattern Recognition
The British Machine Vision Conference is organised by The British Machine Vision Association and Society for Pattern Recognition. The Association is a Company limited by guarantee, No.2543446, and a non-profit-making body, registered in England and Wales as Charity No.1002307 (Registered Office: Dept. of Computer Science, Durham University, South Road, Durham, DH1 3LE, UK).

Imprint | Data Protection