One Loss For All: Deep Hashing With A Single Cosine Similarity Based Learning Objective

Abstract

A deep hashing model typically has two main learning objectives: to make the learned binary hash codes discriminative and to minimize a quantization error. With further constraints such as bit balance and code orthogonality, it is not uncommon for existing models to employ a large number (>4) of losses. This leads to difficulties in model training and subsequently impedes their effectiveness. In this work, we propose a novel deep hashing model with only a single learning objective. Specifically, we show that maximizing the cosine similarity between the continuous codes and their corresponding binary orthogonal codes can ensure both hash code discriminativeness and quantization error minimization. Further, with this learning objective, code balancing can be achieved by simply using a Batch Normalization (BN) layer and multi-label classification is also straightforward with label smoothing. The result is an one-loss deep hashing model that removes all the hassles of tuning the weights of

Stats

• citations0
• S2 citations—
• github stars98
• HF likes0
• heat score3.99
• arxiv keyhoe2021one

Related papers

• Deep Class-wise Hashing: Semantics-preserving Hashing Via Class-wise Loss (2018)11.19
• One Loss For Quantization: Deep Hashing With Discrete Wasserstein Distributional Matching (2022)12.40
• Learning To Hash With Binary Deep Neural Network (2016)14.93
• Compact Hash Code Learning With Binary Deep Neural Network (2017)9.03
• Dual Asymmetric Deep Hashing Learning (2018)9.03
• Semantic Cluster Unary Loss For Efficient Deep Hashing (2018)7.81
• Deep Hashing Via Householder Quantization (2023)0.00
• Deep Priority Hashing (2018)11.67