arXiv Open Access 2021

Contrastive Learning with Large Memory Bank and Negative Embedding Subtraction for Accurate Copy Detection

Shuhei Yokoo

Lihat Sumber

Abstrak

Copy detection, which is a task to determine whether an image is a modified copy of any image in a database, is an unsolved problem. Thus, we addressed copy detection by training convolutional neural networks (CNNs) with contrastive learning. Training with a large memory-bank and hard data augmentation enables the CNNs to obtain more discriminative representation. Our proposed negative embedding subtraction further boosts the copy detection accuracy. Using our methods, we achieved 1st place in the Facebook AI Image Similarity Challenge: Descriptor Track. Our code is publicly available here: \url{https://github.com/lyakaap/ISC21-Descriptor-Track-1st}

Topik & Kata Kunci

cs.CV

Penulis (1)

Shuhei Yokoo

Format Sitasi

APA MLA BibTeX

Yokoo, S. (2021). Contrastive Learning with Large Memory Bank and Negative Embedding Subtraction for Accurate Copy Detection. https://arxiv.org/abs/2112.04323

Akses Cepat

Lihat di Sumber

Informasi Jurnal

Tahun Terbit: 2021
Bahasa: en
Sumber Database: arXiv
Akses: Open Access ✓