Deep Hashing Via Householder Quantization

Abstract

Hashing is at the heart of large-scale image similarity search, and recent methods have been substantially improved through deep learning techniques. Such algorithms typically learn continuous embeddings of the data. To avoid a subsequent costly binarization step, a common solution is to employ loss functions that combine a similarity learning term (to ensure similar images are grouped to nearby embeddings) and a quantization penalty term (to ensure that the embedding entries are close to binarized entries, e.g., -1 or 1). Still, the interaction between these two terms can make learning harder and the embeddings worse. We propose an alternative quantization strategy that decomposes the learning problem in two stages: first, perform similarity learning over the embedding space with no quantization; second, find an optimal orthogonal transformation of the embeddings so each coordinate of the embedding is close to its sign, and then quantize the transformed embedding through the sign func

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyschwengber2023deep

Related papers

• One Loss For Quantization: Deep Hashing With Discrete Wasserstein Distributional Matching (2022)12.40
• Push For Quantization: Deep Fisher Hashing (2019)0.00
• Deep Hashing: A Joint Approach For Image Signature Learning (2016)5.24
• HHF: Hashing-guided Hinge Function For Deep Hashing Retrieval (2021)14.24
• Deep Residual Hashing (2016)0.00
• Simultaneous Compression And Quantization: A Joint Approach For Efficient Unsupervised Hashing (2018)7.50
• Deep Asymmetric Hashing With Dual Semantic Regression And Class Structure Quantization (2021)6.77
• Deep Priority Hashing (2018)11.67