Scalable Person Re-identification on Supervised Smoothed Manifold

Slides:



Advertisements
Similar presentations
January 23 rd, Document classification task We are interested to solve a task of Text Classification, i.e. to automatically assign a given document.
Advertisements

Context-based object-class recognition and retrieval by generalized correlograms by J. Amores, N. Sebe and P. Radeva Discussion led by Qi An Duke University.
Location Recognition Given: A query image A database of images with known locations Two types of approaches: Direct matching: directly match image features.
Foreground Focus: Finding Meaningful Features in Unlabeled Images Yong Jae Lee and Kristen Grauman University of Texas at Austin.
Large-Scale Image Retrieval From Your Sketches Daniel Brooks 1,Loren Lin 2,Yijuan Lu 1 1 Department of Computer Science, Texas State University, TX, USA.
Watching Unlabeled Video Helps Learn New Human Actions from Very Few Labeled Snapshots Chao-Yeh Chen and Kristen Grauman University of Texas at Austin.
Unsupervised Feature Selection for Multi-Cluster Data Deng Cai et al, KDD 2010 Presenter: Yunchao Gong Dept. Computer Science, UNC Chapel Hill.
One-Shot Multi-Set Non-rigid Feature-Spatial Matching
Non-metric affinity propagation for unsupervised image categorization Delbert Dueck and Brendan J. Frey ICCV 2007.
Locally Constraint Support Vector Clustering
Manhattan-world Stereo Y. Furukawa, B. Curless, S. M. Seitz, and R. Szeliski 2009 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.
Nonlinear Dimensionality Reduction Approaches. Dimensionality Reduction The goal: The meaningful low-dimensional structures hidden in their high-dimensional.
A Privacy Preserving Efficient Protocol for Semantic Similarity Join Using Long String Attributes Bilal Hawashin, Farshad Fotouhi Traian Marius Truta Department.
Longbiao Kang, Baotian Hu, Xiangping Wu, Qingcai Chen, and Yan He Intelligent Computing Research Center, School of Computer Science and Technology, Harbin.
Jinhui Tang †, Shuicheng Yan †, Richang Hong †, Guo-Jun Qi ‡, Tat-Seng Chua † † National University of Singapore ‡ University of Illinois at Urbana-Champaign.
Manifold learning: Locally Linear Embedding Jieping Ye Department of Computer Science and Engineering Arizona State University
Discovering Outlier Filtering Rules from Unlabeled Data Author: Kenji Yamanishi & Jun-ichi Takeuchi Advisor: Dr. Hsu Graduate: Chia- Hsien Wu.
A hybrid SOFM-SVR with a filter-based feature selection for stock market forecasting Huang, C. L. & Tsai, C. Y. Expert Systems with Applications 2008.
Paired Sampling in Density-Sensitive Active Learning Pinar Donmez joint work with Jaime G. Carbonell Language Technologies Institute School of Computer.
Query Sensitive Embeddings Vassilis Athitsos, Marios Hadjieleftheriou, George Kollios, Stan Sclaroff.
Institute of Computing Technology, Chinese Academy of Sciences 1 A Unified Framework of Recommending Diverse and Relevant Queries Speaker: Xiaofei Zhu.
Ariadna Quattoni Xavier Carreras An Efficient Projection for l 1,∞ Regularization Michael Collins Trevor Darrell MIT CSAIL.
1 Approximate XML Query Answers Presenter: Hongyu Guo Authors: N. polyzotis, M. Garofalakis, Y. Ioannidis.
Learning Photographic Global Tonal Adjustment with a Database of Input / Output Image Pairs.
Carl Vondrick, Aditya Khosla, Tomasz Malisiewicz, Antonio Torralba Massachusetts Institute of Technology
Advanced Gene Selection Algorithms Designed for Microarray Datasets Limitation of current feature selection methods: –Ignores gene/gene interaction: single.
Ariel Fuxman, Panayiotis Tsaparas, Kannan Achan, Rakesh Agrawal (2008) - Akanksha Saxena 1.
Naifan Zhuang, Jun Ye, Kien A. Hua
Experience Report: System Log Analysis for Anomaly Detection
Outline Introduction State-of-the-art solutions
Deeply learned face representations are sparse, selective, and robust
Guillaume-Alexandre Bilodeau
Intro to Machine Learning
Rule Induction for Classification Using
Supervised Time Series Pattern Discovery through Local Importance
Unsupervised Riemannian Clustering of Probability Density Functions
KDD 2004: Adversarial Classification
Fast Preprocessing for Robust Face Sketch Synthesis
Machine Learning Basics
Huazhong University of Science and Technology
ICCV Hierarchical Part Matching for Fine-Grained Image Classification
mengye ren, ryan kiros, richard s. zemel
COMBINED UNSUPERVISED AND SEMI-SUPERVISED LEARNING FOR DATA CLASSIFICATION Fabricio Aparecido Breve, Daniel Carlos Guimarães Pedronette State University.
Janardhan Rao (Jana) Doppa, Alan Fern, and Prasad Tadepalli
Graph Based Multi-Modality Learning
A Markov Random Field Model for Term Dependencies
Design of Hierarchical Classifiers for Efficient and Accurate Pattern Classification M N S S K Pavan Kumar Advisor : Dr. C. V. Jawahar.
Fine-Grained Visual Categorization
Discriminative Frequent Pattern Analysis for Effective Classification
MEgo2Vec: Embedding Matched Ego Networks for User Alignment Across Social Networks Jing Zhang+, Bo Chen+, Xianming Wang+, Fengmei Jin+, Hong Chen+, Cuiping.
[jws13] Evaluation of instance matching tools: The experience of OAEI
Deep Cross-media Knowledge Transfer
Biased Random Walk based Social Regularization for Word Embeddings
Binghui Wang, Le Zhang, Neil Zhenqiang Gong
Socialized Word Embeddings
Automatic Segmentation of Data Sequences
GANG: Detecting Fraudulent Users in OSNs
Zhedong Zheng, Liang Zheng and Yi Yang
A Classification-based Approach to Question Routing in Community Question Answering Tom Chao Zhou 22, Feb, 2010 Department of Computer.
Related Work in Camera Network Tracking
Extracting Patterns and Relations from the World Wide Web
Rachit Saluja 03/20/2019 Relation Extraction with Matrix Factorization and Universal Schemas Sebastian Riedel, Limin Yao, Andrew.
An Efficient Projection for L1-∞ Regularization
Learning and Memorization
Extracting Why Text Segment from Web Based on Grammar-gram
Learning to Detect Human-Object Interactions with Knowledge
Eliminating Background-Bias for Robust Person Re-identification
Visual Grounding.
Learning Dual Retrieval Module for Semi-supervised Relation Extraction
Machine Learning.
Presentation transcript:

Scalable Person Re-identification on Supervised Smoothed Manifold Song Bai, Xiang Bai, Qi Tian Huazhong University of Science and Technology University of Texas at San Antonio

Task Definition Search a probe in a gallery with the help of the training set Training set Y in pairwise constraints Testing set = Probe p + Gallery X Cam-A Probe Cam-B Gallery Remark: Pairwise-constrainted labels are different from category labels. 2018/4/12 Supervised Smoothed Manifold

Affinity learning in Person Re-identification Most Previous works (Feature + Metric Learning) Our contribution (Supervised Smoothed Manifold) Remark: a new direction which can be investigated. 2018/4/12 Supervised Smoothed Manifold

Motivation The illustration of the feature space True matching pairs A positive pair (labeled) An unlabeled pair A negative pair (labeled) True matching pairs False matching pairs 2018/4/12 Supervised Smoothed Manifold

Supervised Smoothed Manifold Solution On-line Construct the affinity graph with the gallery X and training set Y; Do affinity learning by propagating the pairwise-constrained labels; Return the manifold parameterized by the learning affinity. Off-line Input a given probe image; Do probe embedding on the manifold; Return the matching probabilities. More details can be found in the paper. 2018/4/12 Supervised Smoothed Manifold

Supervised Smoothed Manifold Advantage Three merits: Supervision: take advantage of the supervision in pairwise constraints, which is easily accessible in this task. Efficiency: highly-efficient on-line indexing. Generalization: a post-processing procedure (or a generic tool) to further boost the identification accuracies of most existing algorithms. 2018/4/12 Supervised Smoothed Manifold

Experimental Comparison Datasets: GRID, VIPeR, PRID450S, CUHK03 and Market-1501. Feature: LOMO, GOG, EFL6 and ResNet. Metric: Euclidean and XQDA. Boost the performances of various features and metrics. 2018/4/12 Supervised Smoothed Manifold

Experimental Comparison A new state-of-the-art on five benchmark datasets. Datasets R=1 R=5 R=10 R=20 mAP GRID 24.80 - 58.40 68.88 27.20 61.12 70.56 VIPeR 63.92 91.49 96.65 53.73 96.08 PRID450S 68.40 94.50 97.80 72.98 96.76 99.11 CUHK03 (labeled) 75.3 91.0 96.0 76.6 94.6 98.0 CUHK03 (detected) 65.5 88.4 94.8 72.7 92.4 96.1 Market-1501 76.04 48.45 88.18 76.18 Remark: for each dataset, the 1st row presents the best performances to date, and the 2nd row presents our results. 2018/4/12 Supervised Smoothed Manifold

Supervised Smoothed Manifold Conclusion Addresses person re-identification using affinity learning (or re-ranking, post-ranking), which did not receive enough attention in this field. A simple yet effective tool for various existing algorithms to further improve their performances. 2018/4/12 Supervised Smoothed Manifold

Thanks! Codes are available at https://sites.google.com/site/songbaihust/.