Image similarity using Deep CNN and Curriculum Learning

Abstract

Image similarity involves fetching similar looking images given a reference image. Our solution called SimNet, is a deep siamese network which is trained on pairs of positive and negative images using a novel online pair mining strategy inspired by Curriculum learning. We also created a multi-scale CNN, where the final image embedding is a joint representation of top as well as lower layer embedding's. We go on to show that this multi-scale siamese network is better at capturing fine grained image similarities than traditional CNN's.

9 pages, 6 figures, GHCI 17 conference

References in corpus (4)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Image Matching Using SIFT, SURF, BRIEF and ORB: Performance Comparison for Distorted Images
• Learning to Compare Image Patches via Convolutional Neural Networks
• Recent Advance in Content-based Image Retrieval: A Literature Survey

Cited by in corpus (6)

• OREOS: Oriented Recognition of 3D Point Clouds in Outdoor Scenarios
• 3DEG: Data-Driven Descriptor Extraction for Global re-localization in subterranean environments
• Multi-step feature fusion for natural disaster damage assessment on satellite images
• TACo: Token-aware Cascade Contrastive Learning for Video-Text Alignment
• GetNet: Get Target Area for Image Pairing
• Place recognition survey: An update on deep learning approaches