Computer Example of Schiele and Pentland’s Probabilistic Object Recognition and Localization Marshall Tappen 6.899.

Slides:



Advertisements
Similar presentations
Point Processing Histograms. Histogram Equalization Histogram equalization is a powerful point processing enhancement technique that seeks to optimize.
Advertisements

Michele Merler Jacquilene Jacob.  Applications online are inherently insecure  Growing rate of hackers  Confidentiality of online systems should be.
Improved Census Transforms for Resource-Optimized Stereo Vision
Histograms of Oriented Gradients for Human Detection
Max Cut Problem Daniel Natapov.
Lecture 1: Overview of Computers & Programming
1 Challenge the future HON4D: Histogram of Oriented 4D Normals for Activity Recognition from Depth Sequences Omar Oreifej Zicheng Liu CVPR 2013.
Fitting: The Hough transform. Voting schemes Let each feature vote for all the models that are compatible with it Hopefully the noise features will not.
Mining Distance-Based Outliers in Near Linear Time with Randomization and a Simple Pruning Rule Stephen D. Bay 1 and Mark Schwabacher 2 1 Institute for.
Computer Vision for Human-Computer InteractionResearch Group, Universität Karlsruhe (TH) cv:hci Dr. Edgar Seemann 1 Computer Vision: Histograms of Oriented.
Internet Vision - Lecture 3 Tamara Berg Sept 10. New Lecture Time Mondays 10:00am-12:30pm in 2311 Monday (9/15) we will have a general Computer Vision.
Complex Feature Recognition: A Bayesian Approach for Learning to Recognize Objects by Paul A. Viola Presented By: Emrah Ceyhan Divin Proothi Sherwin Shaidee.
HYBRID-BOOST LEARNING FOR MULTI-POSE FACE DETECTION AND FACIAL EXPRESSION RECOGNITION Hsiuao-Ying ChenChung-Lin Huang Chih-Ming Fu Pattern Recognition,
Instructor: Mircea Nicolescu Lecture 13 CS 485 / 685 Computer Vision.
JPEG2000 Performance Result (so far…) A HiRISE-centric update by Myche McAuley
Lecture 5 Template matching
1 Experimental Methodology H Experimental methods can be used to: – demonstrate that a new concept, technique, or algorithm is feasible –demonstrate that.
Ensemble Tracking Shai Avidan IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE February 2007.
Chapter 9 Virtual Memory Produced by Lemlem Kebede Monday, July 16, 2001.
PCA Channel Student: Fangming JI u Supervisor: Professor Tom Geoden.
September 25, 2014Computer Vision Lecture 6: Spatial Filtering 1 Computing Object Orientation We compute the orientation of an object as the orientation.
Anti-Faces for Detection Daniel Keren Rita Osadchy Haifa University Craig Gotsman Technion Journal Version:
Preprocessing Images for Facial Recognition Adam Schreiner ECE533.
Large-Scale Content-Based Image Retrieval Project Presentation CMPT 880: Large Scale Multimedia Systems and Cloud Computing Under supervision of Dr. Mohamed.
MATLAB Programming Session
:. Standard 3: The Physical Setting ○ Indicator 6.3.3: Explain that the Earth is one of several planets that orbit the sun, and that the moon, as well.
University of Kurdistan Digital Image Processing (DIP) Lecturer: Kaveh Mollazade, Ph.D. Department of Biosystems Engineering, Faculty of Agriculture,
1 Recognition by Appearance Appearance-based recognition is a competing paradigm to features and alignment. No features are extracted! Images are represented.
CSC321: Neural Networks Lecture 2: Learning with linear neurons Geoffrey Hinton.
Generalized Hough Transform
Learning to perceive how hand-written digits were drawn Geoffrey Hinton Canadian Institute for Advanced Research and University of Toronto.
Face Recognition: An Introduction
CVPR 2003 Tutorial Recognition and Matching Based on Local Invariant Features David Lowe Computer Science Department University of British Columbia.
Fitting: The Hough transform
Lecture 7: Features Part 2 CS4670/5670: Computer Vision Noah Snavely.
Virtual Memory Lei Li CS147.
COMP322/S2000/L171 Robot Vision System Major Phases in Robot Vision Systems: A. Data (image) acquisition –Illumination, i.e. lighting consideration –Lenses,
CSC508 Convolution Operators. CSC508 Convolution Arguably the most fundamental operation of computer vision It’s a neighborhood operator –Similar to the.
Histograms of Oriented Gradients for Human Detection(HOG)
Program Design. The design process How do you go about writing a program? –It’s like many other things in life Understand the problem to be solved Develop.
Detecting Image Features: Corner. Corners Given an image, denote the image gradient. C is symmetric with two positive eigenvalues. The eigenvalues give.
Presentation 3 Ruben Villegas Period: 05/31/2012 – 06/03/2012.
Distinctive Image Features from Scale-Invariant Keypoints Presenter :JIA-HONG,DONG Advisor : Yen- Ting, Chen 1 David G. Lowe International Journal of Computer.
1 Image Search/Thinkin g Look at a computer or a photo of a computer. What parts can you identify? 2 Web Search What is hardware? What are three.
Evaluation of Gender Classification Methods with Automatically Detected and Aligned Faces Speaker: Po-Kai Shen Advisor: Tsai-Rong Chang Date: 2010/6/14.
Matrix Multiplication in Hadoop
SIFT.
FACE RECOGNITION. A facial recognition system is a computer application for automatically identifying or verifying a person from a digital image or a.
FINGERTEC FACE ID FACE RECOGNITION Technology Overview.
Computer Vision. Overview of the field  Image / Video => Data  Compare to graphics (the reverse)  Sample applications  Video Camera feed => ID room.
Installing Windows Server 2008
Phases of the Moon.
Distinctive Image Features from Scale-Invariant Keypoints
SIFT paper.
Fitting.
Complexity In examining algorithm efficiency we must understand the idea of complexity Space complexity Time Complexity.
Submitted by: Ala Berawi Sujod Makhlof Samah Hanani Supervisor:
Recovery from Occlusion in Deep Feature Space for Face Recognition
State-of-the-art face recognition systems
A Tutorial on HOG Human Detection
The SIFT (Scale Invariant Feature Transform) Detector and Descriptor
Outline H. Murase, and S. K. Nayar, “Visual learning and recognition of 3-D objects from appearance,” International Journal of Computer Vision, vol. 14,
Oral presentation for ACM International Conference on Multimedia, 2014
Picture Analysis Terms
SIFT.
Grey Level Enhancement
Picture Analysis Terms
Intensity Transformation
Sketch Object Prediction
Anti-Faces for Detection
Presentation transcript:

Computer Example of Schiele and Pentland’s Probabilistic Object Recognition and Localization Marshall Tappen 6.899

Implementation Details You Need to Think About ► How many cells per dimension of the histogram?  It appears that the authors use around 4 or 5 cells per histogram ► I used 6 ► Do all dimensions have the same number of cells? (Yes) ► How do you allocate the cells?  I set the range encompassed by the histogram to cover the full range of possible values ► You could set it to encompass 97% of the values

Comments ► Histograms are big!  720 images (20 objects, 36 views) takes 268 MB of space ► Smaller histograms decrease space needed dramatically ► Executing convolutions to take measurements is by far the slowest part ► Algorithm is very simple to implement  Getting a good training set is the hard part

Testing Procedure ► I used the Columbia Object Image Library (COIL-20) Database ► Grayscale Images of 20 Objects  72 Images of each object, taken at different poses

Testing Procedure ► For each image, I used one half of the poses as the training set ► The other half of the poses were the test set

How Many Bins in Each Histogram? ► My first question was how many bins should be in each dimension of the histogram. ► My second question was how many measurements I need to identify an object.

How Many Bins in Each Histogram?

How Many Views Do We Need? ► My next question was how many views are needed. ► For this test, I chose a portion of the test set and tested it while changing the number of poses in the training set  (I was using 75 measurements and 6 bins per dimension)

How Many Views Do We Need?

How well does it handle varying object size? ► I wanted to know how robust the algorithm is to changing object sizes ► I tested this by using rescaled versions of the test set as input images

How well does it handle varying object size?

What about new lighting? ► I want to see how robust the algorithm is to changing illumination. ► I faked a new illumination by masking the input images ► I masked the test images with a horizontal gradient with a horizontal gradient ► I varied the slope of the gradient the gradient (i.e the beginning value of the gradient)

What about new lighting?

Comments ► The algorithm is surprisingly robust ► It could really benefit from doing the convolutions in hardware ► If you want to use it, ask your advisor to buy more RAM