Filters








2,165 Hits in 5.4 sec

Beyond Product Quantization: Deep Progressive Quantization for Image Retrieval [article]

Lianli Gao, Xiaosu Zhu, Jingkuan Song, Zhou Zhao, Heng Tao Shen
2019 arXiv   pre-print
In this work, we propose a deep progressive quantization (DPQ) model, as an alternative to PQ, for large scale image retrieval.  ...  Product Quantization (PQ) has long been a mainstream for generating an exponentially large codebook at very low memory/time cost.  ...  Acknowledgements This work is supported by the Fundamental Research Funds for the Central Universities (Grant No. ZYGX2014J063, No.  ... 
arXiv:1906.06698v2 fatcat:5onnazjmbnhrnapczrqfcc5p7a

Compressing Deep Convolutional Networks using Vector Quantization [article]

Yunchao Gong and Liu Liu and Ming Yang and Lubomir Bourdev
2014 arXiv   pre-print
Deep convolutional neural networks (CNN) has become the most promising method for object recognition, repeatedly demonstrating record breaking results for image classification and object detection in recent  ...  Simply applying k-means clustering to the weights or conducting product quantization can lead to a very good balance between model size and recognition accuracy.  ...  the gold standard for object recognition, image classification, and retrieval.  ... 
arXiv:1412.6115v1 fatcat:qmfcwljfjjaubmfjw3mxgegn2y

Fixed Point Quantization of Deep Convolutional Networks [article]

Darryl D. Lin, Sachin S. Talathi, V. Sreekanth Annapureddy
2016 arXiv   pre-print
In recent years increasingly complex architectures for deep convolution networks (DCNs) have been proposed to boost the performance on image recognition tasks.  ...  In this paper, we propose a quantizer design for fixed point implementation of DCNs.  ...  Introduction Recent advances in the development of deep convolution networks (DCNs) have led to significant progress in solving non-trivial machine learning problems involving image recognition (Krizhevsky  ... 
arXiv:1511.06393v3 fatcat:v6amoxpaojhm3iparrldbsbjly

PECAN: A Product-Quantized Content Addressable Memory Network [article]

Jie Ran, Rui Lin, Jason Chun Lok Li, Jiajun Zhou, Ngai Wong
2022 arXiv   pre-print
A novel deep neural network (DNN) architecture is proposed wherein the filtering and linear transform are realized solely with product quantization (PQ).  ...  Experiments confirm the feasibility of such Product-Quantized Content Addressable Memory Network (PECAN), which has strong implication on hardware-efficient deployments especially for in-memory computing  ...  Both angle-and distance-based measures are developed for similarity matching of prototypes in product quantization for different complexity-accuracy tradeoffs.  ... 
arXiv:2208.13571v1 fatcat:d75b7235pzcx7neuzpq7hzxqqu

From Hard to Soft: Understanding Deep Network Nonlinearities via Vector Quantization and Statistical Inference [article]

Randall Balestriero, Richard G. Baraniuk
2018 arXiv   pre-print
Nonlinearity is crucial to the performance of a deep (neural) network (DN).  ...  To date there has been little progress understanding the menagerie of available nonlinearities, but recently progress has been made on understanding the rôle played by piecewise affine and convex nonlinearities  ...  INTRODUCTION Deep (neural) networks (DNs) have recently come to the fore in a wide range of machine learning tasks, from regression to classification and beyond.  ... 
arXiv:1810.09274v1 fatcat:tbycafpyjjbkpkuhzcopo2rja4

Efficient Visual Recognition

Li Liu, Matti Pietikäinen, Jie Qin, Wanli Ouyang, Luc Van Gool
2020 International Journal of Computer Vision  
Learning multifunctional binary codes for personal- ized image retrieval Personalized image retrieval A general framework for deep supervised discrete hashing Image retrieval Product quantization  ...  Jin and Junsong Yuan studies the problem of efficient image retrieval and proposes a method called Product Quantization Network (PQN).  ... 
doi:10.1007/s11263-020-01351-w fatcat:mbcq6shmerbo5njayscgb3t4rq

DeepHash: Getting Regularization, Depth and Fine-Tuning Right [article]

Jie Lin, Olivier Morere, Vijay Chandrasekhar, Antoine Veillard, Hanlin Goh
2015 arXiv   pre-print
This work focuses on representing very high-dimensional global image descriptors using very compact 64-1024 bit binary hashes for instance retrieval.  ...  In-depth evaluation shows that our scheme consistently outperforms state-of-the-art methods across all data sets for both Fisher Vectors and Deep Convolutional Neural Network features, by up to 20 percent  ...  DeepHash DeepHash is a hashing scheme based on a deep network to generate binary compact hashes for image instance retrieval ( Figure 1 ). 1 Given a global image descriptor z 0 , a deep network performs  ... 
arXiv:1501.04711v1 fatcat:njzvi55g7rfg5j42obn35wycqq

Face Search at Scale

Dayong Wang, Charles Otto, Anil K. Jain
2017 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Despite significant progress in face recognition, searching a large collection of unconstrained face images remains a difficult problem.  ...  The k retrieved candidates are re-ranked by combining similarities based on deep features and those output by the COTS matcher.  ...  Using product quantization for fast matching based on deep features, we can retrieve the top-k candidate faces in about 0.9 seconds for a 5M image gallery and in about 6.7 seconds for an 80M gallery.  ... 
doi:10.1109/tpami.2016.2582166 pmid:27333599 fatcat:wirvy4vddjafnfna4i4mpkz4n4

Low-Power Computer Vision: Status, Challenges, Opportunities [article]

Sergei Alyamkin, Matthew Ardi, Alexander C. Berg, Achille Brighton, Bo Chen, Yiran Chen, Hsin-Pai Cheng, Zichen Fan, Chen Feng, Bo Fu, Kent Gauen, Abhinav Goel, Alexander Goncharenko (+28 others)
2019 arXiv   pre-print
Computer vision has achieved impressive progress in recent years. Meanwhile, mobile phones have become the primary computing platforms for millions of people.  ...  This article serves two main purposes: (1) Examine the state-of-the-art for low-power solutions to detect objects in images.  ...  For example, people use smartphones for comparison shopping by taking photographs of interested products and search for reviews, similar products, and prices.  ... 
arXiv:1904.07714v1 fatcat:nfss4oa5mfbr7gb34ky2q6fmsq

Content-based Image Retrieval and the Semantic Gap in the Deep Learning Era [article]

Björn Barz, Joachim Denzler
2020 arXiv   pre-print
Content-based image retrieval has seen astonishing progress over the past decade, especially for the task of retrieving images of the same object that is depicted in the query image.  ...  We conclude that the key problem for the further advancement of semantic image retrieval lies in the lack of a standardized task definition and an appropriate benchmark dataset.  ...  [4] created a novel landmarks dataset with over 200,000 images for training purposes, which was later used by other works on deep image retrieval as well [20] .  ... 
arXiv:2011.06490v1 fatcat:fgrcgt2jxbdchfe7ts7t6ephcy

DeepHand: Robust Hand Pose Estimation by Completing a Matrix Imputed with Deep Features

Ayan Sinha, Chiho Choi, Karthik Ramani
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Details of product quantization are skipped for brevity.  ...  retrieval of nearest neighbors from a large population of pre-computed activation features using product quantization. 2.  ... 
doi:10.1109/cvpr.2016.450 dblp:conf/cvpr/SinhaCR16 fatcat:ysuzkkimbrh5ba5v4p3gmevuwe

Challenges and opportunities of image and video retrieval

Guoping Qiu
2022 Frontiers in Imaging  
August CITATION Qiu G ( ) Challenges and opportunities of image and video retrieval. Front. Imaging. : . doi: .  ...  With the development of new technologies such as deep learning, there have already been many works applying deep learning for image retrieval.  ...  Automatic image captioning remains challenging despite impressive progress in deep learning-based solutions.  ... 
doi:10.3389/fimag.2022.951934 fatcat:g4wlnaaxibd57d46uevctnk74e

Sketch-based manga retrieval using manga109 dataset

Yusuke Matsui, Kota Ito, Yuji Aramaki, Azuma Fujimoto, Toru Ogawa, Toshihiko Yamasaki, Kiyoharu Aizawa
2016 Multimedia tools and applications  
It consists of efficient margin labeling, edge orientation histogram feature description, and approximate nearest-neighbor search using product quantization.  ...  To the best of our knowledge, Manga109 is currently the biggest dataset of manga images available for research.  ...  [51] for recent rapid progress in deep architecture.  ... 
doi:10.1007/s11042-016-4020-z fatcat:jjd377iwvrdb3eagtiko7vhima

Transactional Support for Visual Instance Search [chapter]

Herwig Lejsek, Friðrik Heiðar Ásmundsson, Björn Þór Jónsson, Laurent Amsaleg
2018 Lecture Notes in Computer Science  
[32] proposed an indexing scheme based on product quantization that uses ten computers to fit in memory the 1.5 billion images collection they index.  ...  Several variants of product quantization have been published; in particular, Sun et al.  ... 
doi:10.1007/978-3-030-02224-2_6 fatcat:b3h3opoxyndwnetzcui7by2dci

A Survey on Green Deep Learning [article]

Jingjing Xu, Wangchunshu Zhou, Zhiyi Fu, Hao Zhou, Lei Li
2021 arXiv   pre-print
For each category, we discuss the progress that has been achieved and the unresolved challenges.  ...  This paper focuses on presenting a systematic review of the development of Green deep learning technologies.  ...  5) Is linear algebra is the only basic theory for deep learning and whether can we develop a new operation system beyond linear algebra?  ... 
arXiv:2111.05193v2 fatcat:t2blz24y2jakteeeawqqogbkpy
« Previous Showing results 1 — 15 out of 2,165 results