SHAHAB iCV Research Group.

Slides:



Advertisements
Similar presentations
PERCEPTION is the process of organizing and interpreting sensory information.
Advertisements

Object recognition and scene “understanding”
Perception. Rules of Perceptual Organization Gestalt Psychologists The whole is more than the sum of its parts Closure Even if there are gaps in a picture.
Unit 4: Sensation & Perception
Ivan Laptev IRISA/INRIA, Rennes, France September 07, 2006 Boosted Histograms for Improved Object Detection.
Computer Vision for Human-Computer InteractionResearch Group, Universität Karlsruhe (TH) cv:hci Dr. Edgar Seemann 1 Computer Vision: Histograms of Oriented.
Lecture 31: Modern object recognition
My Group’s Current Research on Image Understanding.
Data-driven Visual Similarity for Cross-domain Image Matching
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.
Detecting Pedestrians by Learning Shapelet Features
The Viola/Jones Face Detector (2001)
CPSC 425: Computer Vision (Jan-April 2007) David Lowe Prerequisites: 4 th year ability in CPSC Math 200 (Calculus III) Math 221 (Matrix Algebra: linear.
Visual Cognition II Object Perception. Theories of Object Recognition Template matching models Feature matching Models Recognition-by-components Configural.
A Study of Approaches for Object Recognition
Visual Cognition II Object Perception. Theories of Object Recognition Template matching models Feature matching Models Recognition-by-components Configural.
Color, lightness & brightness Lavanya Sharan February 7, 2011.
Con-Text: Text Detection Using Background Connectivity for Fine-Grained Object Classification Sezer Karaoglu, Jan van Gemert, Theo Gevers 1.
Lecture 29: Recent work in recognition CS4670: Computer Vision Noah Snavely.
Computer vision.
The Three R’s of Vision Jitendra Malik.
Final Exam Review CS485/685 Computer Vision Prof. Bebis.
Internet-scale Imagery for Graphics and Vision James Hays cs195g Computational Photography Brown University, Spring 2010.
Introduction to Computer Vision Olac Fuentes Computer Science Department University of Texas at El Paso El Paso, TX, U.S.A.
Object Tracking/Recognition using Invariant Local Features Applications l Mobile robots, driver assistance l Cell phone location or object recognition.
Why Categorize in Computer Vision ?. Why Use Categories? People love categories!
Recovering Surface Layout from a Single Image D. Hoiem, A.A. Efros, M. Hebert Robotics Institute, CMU Presenter: Derek Hoiem CS 598, Spring 2009 Jan 29,
Computer Science Department Pacific University Artificial Intelligence -- Computer Vision.
Reading Between The Lines: Object Localization Using Implicit Cues from Image Tags Sung Ju Hwang and Kristen Grauman University of Texas at Austin Jingnan.
Lecture 31: Modern recognition CS4670 / 5670: Computer Vision Noah Snavely.
80 million tiny images: a large dataset for non-parametric object and scene recognition CS 4763 Multimedia Systems Spring 2008.
Pedestrian Detection and Localization
MSRI workshop, January 2005 Object Recognition Collected databases of objects on uniform background (no occlusions, no clutter) Mostly focus on viewpoint.
Lecture 7: Features Part 2 CS4670/5670: Computer Vision Noah Snavely.
Histograms of Oriented Gradients for Human Detection(HOG)
Jack Pinches INFO410 & INFO350 S INFORMATION SCIENCE Computer Vision I.
Robotics and Computer Vision
Improved Object Detection
Duc-Tien Dang-Nguyen, Giulia Boato, Alessandro Moschitti, Francesco G.B. De Natale Department to Information and Computer Science –University of Trento.
Finding Clusters within a Class to Improve Classification Accuracy Literature Survey Yong Jae Lee 3/6/08.
To Students in Psychology Some of you have used part of at least one Trinity Day learning about color perception on your own. The internet is.
Week 4: 6/6 – 6/10 Jeffrey Loppert. This week.. Coded a Histogram of Oriented Gradients (HOG) Feature Extractor Extracted features from positive and negative.
AN ACTIVE VISION APPROACH TO OBJECT SEGMENTATION – Paul Fitzpatrick – MIT CSAIL.
Does one size really fit all? Evaluating classifiers in a Bag-of-Visual-Words classification Christian Hentschel, Harald Sack Hasso Plattner Institute.
Visual Information Processing. Human Perception V.S. Machine Perception  Human perception: pictorial information improvement for human interpretation.
- photometric aspects of image formation gray level images
Object detection with deformable part-based models
Presented by David Lee 3/20/2006
Perception The process of organizing and interpreting information, enabling us to recognize meaningful objects and events.
Object Recognition Cognitive Robotics David S. Touretzky &
R. C. James Photograph.
THE VISUAL SYSTEM: PERCEPTUAL PROCESSES
Feature description and matching
UGS 302 Usability and User-Centered Design
Object detection as supervised classification
Introduction of Pedestrian Detection
Cheng-Ming Huang, Wen-Hung Liao Department of Computer Science
A Tutorial on HOG Human Detection
Features Readings All is Vanity, by C. Allan Gilbert,
Computer Vision Computer vision attempts to construct meaningful and explicit descriptions of the world depicted in an image Using machines to Interpret!!!
Aim of the project Take your image Submit it to the search engine
Brief Review of Recognition + Context
Large Scale Image Deduplication
KFC: Keypoints, Features and Correspondences
CMSC 426: Image Processing (Computer Vision)
Perceptual Organization
Feature descriptors and matching
Lecture: Object Recognition
Learning complex visual concepts
THE ASSISTIVE SYSTEM SHIFALI KUMAR BISHWO GURUNG JAMES CHOU
Presentation transcript:

SHAHAB iCV Research Group

Image Processing Prospective

IP/IU Wikipedia: image processing is any form of signal processing for which the input is an image. Any kind of image manipulation is referred as image processing.

IP/IU Vision is the process of discovering from images what is present in the world, and where it is. -- David Marr, Vision (1982) Humans can perceive and interpret images very fast and accurately.

Image Formation Scene Lens Shutter Area CCD 46 44 47 50 56 40 48 76 62 69 45 36 39 88 87 65 38 90 41 43 112 77 66 72 108 74 98 86 91 83 57

What Do We See? 3D world 2D image Figures © Stephen E. Palmer, 2002 This emphasizes that interpretation is inherently a difficult problem. It is ill-posed because there an infinite number of interpretations of a 3D scene. Figures © Stephen E. Palmer, 2002

What do we see? 3D world 2D image Painted backdrop It is impossible to know from pure visual processing if this scene is a backdrop or a real 3D scene. Painted backdrop Figures © Stephen E. Palmer, 2002

Brightness: Measurement vs. Perception Visual system tries to undo the measured brightness into the reflectance and illumination and estimate the reflectance that is inherent to the object. Do squares A and B have the same brightness?

Brightness: Measurement vs. Perception Squares A and B have the same measured brightness but a different perceived brightness!

Lengths: Measurement vs. Perception Müller-Lyer Illusion Our perception of geometric properties is affected by our interpretation. http://www.michaelbach.de/ot/sze_muelue/index.html

Which monster is larger? Shepard RN (1990) Mind Sights: Original Visual Illusions, Ambiguities, and other Anomalies, New York: WH Freeman and Company We can’t help but to integrate perspective cues into our interpretation of the image.

Find The Face In This “Coffee Beans” Image? We ourselves are susceptible to clutter as well. This is a problem where computer might do faster than human. http://www.michaelbach.de/ot/sze_muelue/index.html

Brain Fills In Occlusions

Segmentation and Grouping – Find The Dog This scene is really a collection of many random spots and dots but the brain can group them together to segment the dog from the background.

Is Face Processing Orientation Dependent?

What is context? Context is information relevant to the detection task but not directly due to the physical appearance of the object. Wolf and Bileschi, 2004 What do we mean by “context”? We can interpret the H or A based on context. Example from “Cognition in Action” Smyth Collins Morris Levy, 1994, LEA Publishers. In our case, we think of “the object” as a person’s face. Smyth et al., 1994

Role of Context in Image Understanding TENNIS BALL Inter-object semantics [Rabinovich 2007]

Examples Pedestrian detection Face detection Object recognition Geolocating images from content Event recognition using ground and aerial images Annotation of pictures

General Object Recognition Problem: Find many different types of general objects in images Feature: Histogram of Orientated Gradients (HOG) Dataset: PASCAL Classifier: SVM + Graphical Model Performance: Avg. Precision: horse 32%, person 42%, sofa 14%, car 33% On Inria Person Dataset: 86.9% avg. precision Image First feature selected by adaboost Second feature selected by adaboost

Event Recognition Using Ground and Aerial Images Problem: Find the event in an image from visual content and GPS coordinates Feature: Scale Invariant Feature Transform (SIFT) and Color Moments Dataset: Geo-tagged images from 1) Internet, and 2) Community collections Classifier: SVM + Boosted Trees Accuracy: theme-park 90%, forest 84%, beach 80%, city 58%, tennis 54% Image First feature selected by adaboost Second feature selected by adaboost

Annotation of Pictures Problem: Use visual content to annotate an image using a vocabulary Feature: Daubechies wavelet coefficients Dataset: Corel dataset + 54,700 images from Flickr Classifier: Generalized mixture modeling + D2 clustering Performance: Percentage of images correctly annotated by at least: 1 word 51%, 2 words 65%, 3 words 75% Image First feature selected by adaboost Second feature selected by adaboost

Thank You