Counting in High-Density Crowd Videos

Slides:



Advertisements
Similar presentations
Proposal for a new design of LumiCal R. Ingbir, P. Ruzicka, V. Vrba October 07 Malá Skála.
Advertisements

Early Math Counting & Skip Counting. Early Math “0” – see the number Counting & Skip Counting.
SE 501 Software Development Processes Dr. Basit Qureshi College of Computer Science and Information Systems Prince Sultan University Lecture for Week 7.
Insect Electrophysiology and Training Bee hive entrance/exit counting using DVS Brian H. Smith, Robyn Verrinder and Tobi Delbruck.
ACM Multimedia 2008 Feng Liu 1, Yuhen-Hu 1,2 and Michael Gleicher 1.
Object Detection and Tracking Mike Knowles 11 th January 2005
ENTERFACE ’10 Amsterdam, July-August 2010 Hamdi Dibeklio ğ lu Ilkka Kosunen Marcos Ortega Albert Ali Salah Petr Zuzánek.
Trinity College Dublin PixelGT: A new Ground Truth specification for video surveillance Dr. Kenneth Dawson-Howe, Graphics, Vision and Visualisation Group.
Modeling and Prediction of Abdominal Tumor Motion Haobing Wang Department of Computer Science May 9 th, 2003.
Editing You have many ways to put your video together.
Name: Date: How to Find the Number You Are Skip Counting By 1) Find 2 numbers next to each other. 2) Find the smaller number on the number grid. 3) Hop.
1 The Earth’s Shells Quantitative Concepts and Skills Weighted average The nature of a constraint Volume of spherical shells Concept that an integral is.
computer
Soccer Video Analysis EE 368: Spring 2012 Kevin Cheng.
Logan Lebanoff Mentor: Haroon Idrees. Two-layer method  Trying a method that will have two layers of neural networks.
Dynamic Captioning: Video Accessibility Enhancement for Hearing Impairment Richang Hong, Meng Wang, Mengdi Xuy Shuicheng Yany and Tat-Seng Chua School.
There’s a right way and a wrong way to document data manually and from DAS at the same time … If you enter the data in the wrong sequence, you risk losing.
Crowd Analysis at Mass Transit Sites Prahlad Kilambi, Osama Masound, and Nikolaos Papanikolopoulos University of Minnesota Proceedings of IEEE ITSC 2006.
NCCA STEP UP Challenge HOW TO JOIN THE CONTEST AND LOG YOUR STEPS.
This material is approved for public release. Distribution is limited by the Software Engineering Institute to attendees. Sponsored by the U.S. Department.
Professor: Dr. Shu-Ching Chen TA: Hsin-Yu Ha Stored Procedure used in PosgreSQL.
An Effective & Interactive Approach to Particle Tracking for DNA Melting Curve Analysis 李穎忠 DEPARTMENT OF COMPUTER SCIENCE & INFORMATION ENGINEERING NATIONAL.
Active Frame Selection for Label Propagation in Videos Sudheendra Vijayanarasimhan and Kristen Grauman Department of Computer Science, University of Texas.
Tracking Groups of People for Video Surveillance Xinzhen(Elaine) Wang Advisor: Dr.Longin Latecki.
Quality Assessment Recognition Tasks (QART) – Recent Results Mikołaj Leszczuk, Lucjan Janowski, Łukasz Dudek, Sergio Garcia AGH – University of Science.
Logan Lebanoff Mentor: Haroon Idrees
CTS130 Spreadsheet Lesson 6 Working with Math & Trig, Statistical, and Date & Time Functions.
Detecting Occlusion from Color Information to Improve Visual Tracking
Bayesian Decision Theory Case Studies CS479/679 Pattern Recognition Dr. George Bebis.
Scientific Notation *Used to describe numbers with the powers of 10. Used mainly with very large numbers and very small numbers. Large numbers (greater.
Mental Subtraction – Objectives
Add and Subtract Negative Numbers
OptiSystem applications: SER & BER analysis of QAM-PSK-PAM systems
How to do a “Stop Motion Video” in 2 minutes
OptiSystem applications: BER analysis of BPSK with RS encoding
Tracking parameter optimization
Counting In High Density Crowd Videos
Vehicle Segmentation and Tracking in the Presence of Occlusions
Counting in High-Density Crowd Videos
Data Summarization First, we need to have our database in good shape before going into final analysis Preliminary analysis, however, helps us detect unsuspected.
Counting in High-Density Crowd Videos
Marked Point Processes for Crowd Counting
Counting in Dense Crowds using Deep Learning
Counting & Comparing Money 2 $ $ $ $.
Counting & Comparing Money $ $ $ $.
Online Graph-Based Tracking
VMorph: Motion and Feature-Based Video Metamorphosis
Mentor: Salman Khokhar
DRAMA TECHNIQUES Learning Mat
Hundreds board counting
Segmentation of cardiac MRI using particle filters
CS150 Introduction to Computer Science 1
English 2 - May 23rd Agenda: Grade Check and Reflection
Quiz Name Here Click to start.
Week 1 Alan Wright - UCF.
Project 2: IoT Device Vulnerabilities and Security REU Student: Bryan Pearson Graduate mentors: Kelvin Ly and Kaveh Shamsi Faculty Mentor(s): Dr. Jin.
2017 Year in Review & Building on 2018
Rate of Change The rate of change is the change in y-values over the change in x-values.
Undoing Operations Objective: Students will be able to identify what others have done wrong in a problem and work things backwards.
We are learning to … Count in 3s!
2018 Year in Review & Building on 2019
By Claire Barnes, Willow Dene School
Project Design and Framing
Chapter 12 Historical forecasting techniques
UCF Computer Vision REU 2012 Paul Finkel 6/25/12
Types of Errors And Error Analysis.
Ivette Carreras Haroon Idrees
The return Statement © 2018 Kris Jordan.
Counting in High-Density Crowd Videos
Programming Techniques
Presentation transcript:

Counting in High-Density Crowd Videos Edgar Lopez Mentor: Dr. Haroon Idrees

Video Annotations Matlab script to automatically annotate frames by tracking the points of the manually annotated frames. Method Compute forward tracks. Compute backward tracks. Match forward and backward tracks. Compute average tracks.

Results Every 5th frame was manually annotated.

Results Every 30th frame was manually annotated.

Results Every 50th frame was manually annotated.

Challenges People exit and enter the scene. Some points that may be annotated in one frame are not annotated in the next annotated frame. Tracker loses several points between frames. Experiment with different parameters. Alternatively, try with a different tracker. Some tracks are so close to each other that matching is done wrong. The greater the number of frames between manually annotated frames, the larger the error. Add a weighting system to tracks.