Deep Learning to Ternary Hash Codes by Continuation

07/16/2021
by   Mingrui Chen, et al.
0

Recently, it has been observed that 0,1,-1-ternary codes which are simply generated from deep features by hard thresholding, tend to outperform -1,1-binary codes in image retrieval. To obtain better ternary codes, we for the first time propose to jointly learn the features with the codes by appending a smoothed function to the networks. During training, the function could evolve into a non-smoothed ternary function by a continuation method. The method circumvents the difficulty of directly training discrete functions and reduces the quantization errors of ternary codes. Experiments show that the generated codes indeed could achieve higher retrieval accuracy.

READ FULL TEXT

page 1

page 2

research
11/25/2019

Online Hashing with Efficient Updating of Binary Codes

Online hashing methods are efficient in learning the hash functions from...
research
02/01/2019

Deep Triplet Quantization

Deep hashing establishes efficient and effective image retrieval by end-...
research
02/02/2017

HashNet: Deep Learning to Hash by Continuation

Learning to hash has been widely applied to approximate nearest neighbor...
research
01/21/2015

Optimizing affinity-based binary hashing using auxiliary coordinates

In supervised binary hashing, one wants to learn a function that maps a ...
research
06/16/2019

Deep Recurrent Quantization for Generating Sequential Binary Codes

Quantization has been an effective technology in ANN (approximate neares...
research
06/02/2021

LLC: Accurate, Multi-purpose Learnt Low-dimensional Binary Codes

Learning binary representations of instances and classes is a classical ...
research
12/08/2017

Compact Hash Code Learning with Binary Deep Neural Network

In this work, we firstly propose deep network models and learning algori...

Please sign up or login with your details

Forgot password? Click here to reset