Metrik Öǧrenme Tabanli Ürün Tanima ve Yeni Ürün Keşfetme Sistemi

Translated title of the contribution: A Metric Learning Based System for Retail Product Recognition and Novel Class Discovery

Ibrahim Şamil Yalçiner, Zubeyir Genç, Lütfü Çakil, Hazim Kemal Ekenel

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Developing a retail product recognition system presents different challenges compared to traditional object recognition systems. Retail products contain classes that are very similar to each other but differ in small details. The current product classes in the system must be frequently expanded with new product classes or updated with packaging changes. For these reasons, we developed a metric learning-based product recognition system that provides superior product recognition performance and the ability to add new products and update existing ones. The key components of our system are the ConvNext-nano feature extractor, ArcFace loss, feature vector database, similarity search, and clustering algorithm. Through tests performed on a custom dataset consisting of 775,000 samples from 994 classes, we achieved an accuracy of 93.7% in product recognition and an average clustering accuracy of 67.58% in discovering new products.

Translated title of the contributionA Metric Learning Based System for Retail Product Recognition and Novel Class Discovery
Original languageTurkish
Title of host publication31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350343557
DOIs
Publication statusPublished - 2023
Event31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023 - Istanbul, Turkey
Duration: 5 Jul 20238 Jul 2023

Publication series

Name31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023

Conference

Conference31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023
Country/TerritoryTurkey
CityIstanbul
Period5/07/238/07/23

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Fingerprint

Dive into the research topics of 'A Metric Learning Based System for Retail Product Recognition and Novel Class Discovery'. Together they form a unique fingerprint.

Cite this