Human Action Recognition Week 8

Slides:

Advertisements

Similar presentations

Food Recognition Using Statistics of Pairwise Local Features Shulin (Lynn) Yang University of Washington Mei Chen Intel Labs Pittsburgh Dean Pomerleau.

Advertisements

1 Semi-supervised learning for protein classification Brian R. King Chittibabu Guda, Ph.D. Department of Computer Science University at Albany, SUNY Gen*NY*sis.

Human Action Recognition across Datasets by Foreground-weighted Histogram Decomposition Waqas Sultani, Imran Saleemi CVPR 2014.

What’s Your Guess? Chapter 9: Review of Convergent or Divergent Series.

CS395: Visual Recognition Spatial Pyramid Matching Heath Vinicombe The University of Texas at Austin 21 st September 2012.

1 Week 6 Fatemeh Yazdiananari. 2 Feature Extraction 75 Validation Videos Were sent to the cluster for DTF feature extraction The extracted features are.

Addressing the Medical Image Annotation Task using visual words representation Uri Avni, Tel Aviv University, Israel Hayit GreenspanTel Aviv University,

Robust Multi-Kernel Classification of Uncertain and Imbalanced Data

CS771 Machine Learning : Tools, Techniques & Application Gaurav Krishna Y Harshit Maheshwari Pulkit Jain Sayantan Marik

Distributional Clustering of Words for Text Classification L. Douglas Baker Andrew Kachites McCallum SIGIR’98.

Learning kernel matrix by matrix exponential update Jun Liao.

5/30/2006EE 148, Spring Visual Categorization with Bags of Keypoints Gabriella Csurka Christopher R. Dance Lixin Fan Jutta Willamowski Cedric Bray.

Histograms and Polygons Problem 24 page 42. Here I copied the data created in the other section. Notice how I put the labels. I put Frequency instead.

SVMLight SVMLight is an implementation of Support Vector Machine (SVM) in C. Download source from :

Project 2 SIFT Matching by Hierarchical K-means Quantization

Bag of Video-Words Video Representation

Taylor Rassmann.  Look at a confusion matrix of the UCF50 dataset  Dollar Features  Find the two most confused classes  Train an SVM specifically.

Taylor Rassmann.  Almost done extracting six kinematic features from optical flow  Divergence  Vorticity  Symmetric Flow Fields (u and v components)

A Statistical Approach to Speed Up Ranking/Re-Ranking Hong-Ming Chen Advisor: Professor Shih-Fu Chang.

Svetlana Lazebnik, Cordelia Schmid, Jean Ponce

Week5 Amari Lewis Aidean Sharghi. Testing the data for classification Divide the date into TEST and TRAIN data. First the regular.jpeg images Then, the.

Representations for object class recognition David Lowe Department of Computer Science University of British Columbia Vancouver, Canada Sept. 21, 2006.

Combining multiple learners Usman Roshan. Bagging Randomly sample training data Determine classifier C i on sampled data Goto step 1 and repeat m times.

Taylor Rassmann.  Optical flow is still being attained from the UCF50 dataset.  On action 35 out of 50  Currently at 900 GB of data.

Taylor Rassmann.  Grouping data objects into X tree of clusters and uses distance matrices as clustering criteria  Two Hierarchical Clustering Categories:

Mingyang Zhu, Huaijiang Sun, Zhigang Deng Quaternion Space Sparse Decomposition for Motion Compression and Retrieval SCA 2012.

Taylor Rassmann.  Student at the University of Central Florida  Burnett Honors College  Fall 2008 – Fall 2011  Major: Computer Science  Upper Division.

divergent 2.absolutely convergent 3.conditionally convergent.

Week8 Fatemeh Yazdiananari.  Fixed the issues with classifiers  We retrained SVMs with the new UCF101 histograms  On temporally untrimmed videos: ◦

Taylor Rassmann.  Look at a confusion matrix of the UCF50 dataset  Dollar Features  Idea of a structured tree SVM  Different configurations have different.

Musical Genre Categorization Using Support Vector Machines Shu Wang.

Cell Segmentation in Microscopy Imagery Using a Bag of Local Bayesian Classifiers Zhaozheng Yin RI/CMU, Fall 2009.

Next, this study employed SVM to classify the emotion label for each EEG segment. The basic idea is to project input data onto a higher dimensional feature.

Finding Clusters within a Class to Improve Classification Accuracy Literature Survey Yong Jae Lee 3/6/08.

1 Chapter 9. 2 Does converge or diverge and why?

Date of download: 6/29/2016 Copyright © 2016 SPIE. All rights reserved. Overview of the sparse correlaton model for land-use scene classification. Figure.

Clustering (1) Clustering Similarity measure Hierarchical clustering

Antoine Guitton, Geophysics Department, CSM

SLAQ: Quality-Driven Scheduling for Distributed Machine Learning

Human Action Recognition Week 10

Bag-of-Visual-Words Based Feature Extraction

Data Driven Attributes for Action Detection

Temporal Order-Preserving Dynamic Quantization for Human Action Recognition from Multimodal Sensor Streams Jun Ye Kai Li Guo-Jun Qi Kien.

Introductory Seminar on Research: Fall 2017

Exemplar-SVM for Action Recognition

Outline Multilinear Analysis

Using Transductive SVMs for Object Classification in Images

Machine Learning Week 1.

Action Recognition in Temporally Untrimmed Videos

SEG 4630 E-Commerce Data Mining — Final Review —

Revision (Part II) Ke Chen

Ratio Test THE RATIO AND ROOT TESTS Series Tests Test for Divergence

Test the series for convergence or divergence. {image}

Comparison Tests Lesson 9.4.

The Open World of Micro-Videos

REU Week 1 Ivette Carreras UCF.

Test the series for convergence or divergence. {image}

Revision (Part II) Ke Chen

Mentor: Salman Khokhar

Data Driven Attributes for Action Detection

Week 6 Fatemeh Yazdiananari.

Weakly Supervised Action Recognition

Label Name Label Name Label Name Label Name Label Name Label Name

Generalized Locality Preserving Projections

If x is a variable, then an infinite series of the form

Leverage Consensus Partition for Domain-Specific Entity Coreference

CSSE463: Image Recognition Day 11

Chapter 9.4 COMPARISONS OF SERIES.

Multiple Organ detection in CT Volumes Using Random Forests - Week 9

THE ASSISTIVE SYSTEM SHIFALI KUMAR BISHWO GURUNG JAMES CHOU

Presentation transcript:

Human Action Recognition Week 8 Taylor Rassmann Human Action Recognition Week 8

Bag of Words Method tested on first 11 actions of UCF50 dataset Used built in K-Means 500 Centers Kinematic Features

Results: Kinematic Features Average accuracies between 70-90 percent Vorticity Symmetric Flow U

Results: Kinematic Features Asymmetric Flow U Asymmetric Flow V

Hierarchical SVM Use K-Means clustering of labels after codebook and histogram generation Label 1 Label 5 Label 2 Label 8 Label 11 Label 9 Label 3 Label 4 Label 7

Results: Dollar Features 50 Actions Average Accuracies: No centers: 70% 3 centers: 68% 5 centers: 66% 10 centers: 61%

Results: Dollar Features 11 Actions Average Accuracies: No centers: 87% 3 centers: 85%

Confusion Matrix Comparison

Results: Divergence 11 Actions Average Accuracies: No centers: 46% 3 centers: 46%

Confusion Matrix Comparison

Hierarchical SVM Results similar to non-clustering Labels centers converging on one another Label 1 Label 5 Label 2 Label 3 Label 4 Label 7 Label 8 Label 11 Label 9

Current Work Finish K-Means on all 50 kinematic features Divergence is done Make Histograms Test average accuracies with SVMs Find new direction of classification if accuracies are lower than standard SVM