Week 1 Alan Wright - UCF.

Slides:



Advertisements
Similar presentations
A Robust Super Resolution Method for Images of 3D Scenes Pablo L. Sala Department of Computer Science University of Toronto.
Advertisements

3D Model Matching with Viewpoint-Invariant Patches(VIP) Reporter :鄒嘉恆 Date : 10/06/2009.
Motion.
The fundamental matrix F
Analysis of Contour Motions Ce Liu William T. Freeman Edward H. Adelson Computer Science and Artificial Intelligence Laboratory Massachusetts Institute.
MIT AI Lab Paul Viola NTT Visit: Image Database Retrieval Variable Viewpoint Reality Professor Paul Viola Collaborators: Professor Eric Grimson, Jeremy.
Distinctive Image Features from Scale- Invariant Keypoints Mohammad-Amin Ahantab Technische Universität München, Germany.
Large-Scale, Real-World Face Recognition in Movie Trailers Week 2-3 Alan Wright (Facial Recog. pictures taken from Enrique Gortez)
Computer Vision REU Week 2 Adam Kavanaugh. Video Canny Put canny into a loop in order to process multiple frames of a video sequence Put canny into a.
Computer Vision Optical Flow
Crowd Size Estimation Luis Huang ECE 172A - UCSD.
Motion from image and inertial measurements (additional slides) Dennis Strelow Carnegie Mellon University.
Computing motion between images
1 Stanford CS223B Computer Vision, Winter 2006 Lecture 7 Optical Flow Professor Sebastian Thrun CAs: Dan Maynes-Aminzade, Mitul Saha, Greg Corrado Slides.
CSCE 641 Computer Graphics: Image-based Modeling (Cont.) Jinxiang Chai.
A Brief Overview of Computer Vision Jinxiang Chai.
Computer vision.
05 - Feature Detection Overview Feature Detection –Intensity Extrema –Blob Detection –Corner Detection Feature Descriptors Feature Matching Conclusion.
Prakash Chockalingam Clemson University Non-Rigid Multi-Modal Object Tracking Using Gaussian Mixture Models Committee Members Dr Stan Birchfield (chair)
Tzu ming Su Advisor : S.J.Wang MOTION DETAIL PRESERVING OPTICAL FLOW ESTIMATION 2013/1/28 L. Xu, J. Jia, and Y. Matsushita. Motion detail preserving optical.
Detecting Pedestrians Using Patterns of Motion and Appearance Paul Viola Microsoft Research Irfan Ullah Dept. of Info. and Comm. Engr. Myongji University.
UCF Computer Vision REU 2012 Week 1 Presentation Paul Finkel 5/21/12.
CSCE 643 Computer Vision: Structure from Motion
Classifying Images with Visual/Textual Cues By Steven Kappes and Yan Cao.
UCF REU: Weeks 1 & 2. Gradient Code Gradient Direction of the Gradient: Calculating theta.
Computer Vision Spring ,-685 Instructor: S. Narasimhan Wean 5403 T-R 3:00pm – 4:20pm Lecture #16.
Effective Optical Flow Estimation
Lecture 7: Features Part 2 CS4670/5670: Computer Vision Noah Snavely.
Pyramidal Implementation of Lucas Kanade Feature Tracker Jia Huang Xiaoyan Liu Han Xin Yizhen Tan.
1 Motion estimation from image and inertial measurements Dennis Strelow and Sanjiv Singh.
Kylie Gorman WEEK 1-2 REVIEW. CONVERTING AN IMAGE FROM RGB TO HSV AND DISPLAY CHANNELS.
Joint Tracking of Features and Edges STAN BIRCHFIELD AND SHRINIVAS PUNDLIK CLEMSON UNIVERSITY ABSTRACT LUCAS-KANADE AND HORN-SCHUNCK JOINT TRACKING OF.
Spatiotemporal Saliency Map of a Video Sequence in FPGA hardware David Boland Acknowledgements: Professor Peter Cheung Mr Yang Liu.
Motion Estimation Today’s Readings Trucco & Verri, 8.3 – 8.4 (skip 8.3.3, read only top half of p. 199) Newton's method Wikpedia page
PRESENTATION REU IN COMPUTER VISION 2014 AMARI LEWIS CRCV UNIVERSITY OF CENTRAL FLORIDA.
IEEE International Conference on Multimedia and Expo.
SIFT DESCRIPTOR K Wasif Mrityunjay
Motion Estimation Today’s Readings Trucco & Verri, 8.3 – 8.4 (skip 8.3.3, read only top half of p. 199) Newton's method Wikpedia page
Dr. J. Shanbehzadeh M.HosseinKord Science and Research Branch of Islamic Azad University Machine Vision 1/49 slides.
Midterm Review. Tuesday, November 3 7:15 – 9:15 p.m. in room 113 Psychology Closed book One 8.5” x 11” sheet of notes on both sides allowed Bring a calculator.
1 Long-term image-based motion estimation Dennis Strelow and Sanjiv Singh.
REU Week 1 Presented by Christina Peterson. Edge Detection Sobel ◦ Convolve image with derivative masks:  x:  y: ◦ Calculate gradient magnitude ◦ Apply.
Distinctive Image Features from Scale-Invariant Keypoints Presenter :JIA-HONG,DONG Advisor : Yen- Ting, Chen 1 David G. Lowe International Journal of Computer.
Another Example: Circle Detection
Feature Tracking and Optical Flow
Human Action Recognition Week 10
Horn-Schunk Optical Flow
Motion and Optical Flow
Think What will be the output?
Introduction of Real-Time Image Processing
CSE 577 Image and Video Analysis
Report 1: Optical Flow and Sift
Jeremy Bolton, PhD Assistant Teaching Professor
Lecture 26: Faces and probabilities
Computer Vision Lecture 16: Texture II
What I learned in the first 2 weeks
Applications of Cellular Neural Networks to Image Understanding
REU Week 1 Ivette Carreras UCF.
Week 7 Nicholas Baker.
Week 6 Nicholas Baker.
Announcements more panorama slots available now
Analysis of Contour Motions
Announcements more panorama slots available now
Computational Photography
Feature descriptors and matching
Open Topics.
Optical flow and keypoint tracking
Week 1 Emily Hand UNR.
REU Week 1 Jared Rhizor.
Project Presentation – Week 6
Presentation transcript:

Week 1 Alan Wright - UCF

Optical Flow - Lucas Kanade Input images:

Optical Flow - Lucas Kanade Output:

Optical Flow - Lucas Kanade Output of Flow to Color:

Optical Flow - Lucas Kanade Output of both:

Optical Flow - Lucas Kanade Video Script inputPath = '..\Sequence'; for i = 133:2:183 I = [inputPath '\' sprintf('%04d.jpeg', i)]; I2 = [inputPath '\' sprintf('%04d.jpeg', i+1)]; opticflow(I, I2); print('-djpeg', sprintf('%04d.jpg', i) ); close all; end This is your function!

Optical Flow - Lucas Kanade Video Script

Optical Flow - Lucas Kanade Video Script Output

Optical Flow - Lucas Kanade w/ Pyramids Hardcoded 3 levels Changed to variable length of level using MATLab loop. Difficulties encountered: Resolution must be evenly divisible by 2, for each level. Ex: 200x200 lvl1 100x100 lvl2 50 x 50 lvl3 25 x 25 lvl4 12.5 x 12.5 lvl5 Lose a pixel!

Level 3 Level 2 Level 1

Optical Flow: Pyramids vs Non

Optical Flow: Pyramids vs Non With Pyramids: more uniform motion, more accurate, takes more computational time. Incorrect motion exaggerated? Currently using a 3x3 square, may need to be larger for more accurate results. Fix: If the shift is outside the 3x3, do not calculate!

SIFT Descriptor Input: 18x18 patch

SIFT Descriptor 2 pixel border for a 16x16 patch. Split into 16, 4x4 squares. Based on each pixels direction, it's magnitude is placed in a bin. My output: 16x8 matrix. 16 squares, each with 8 bins (0-44, 45-89,...,320-359) Discrepancies: Different sigma values for gaussian blur, center pixel for VLFEAT.

Sift Descriptor

Possible Projects? Large Scale Real World Face Recognition in Movie Trailers - Enrique Ortiz Web Assisted Object Detection for Outdoor Scenes - Amir R. Zamir Crowd counting by estimation of texture repetition - Imran Saleemi 3D Joint Localization for Gesture Recognition