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

Code

• github.com/lyakaap/isc21-descriptor-track-1st
  OfficialIn paperpytorch★ 146
• github.com/towhee-io/towhee
  pytorch★ 3,459

Abstract

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: https://github.com/lyakaap/ISC21-Descriptor-Track-1st

Tasks

Reproductions