ICG Professor Horst Cerjak, 19.12.2005 1 Horst Bischof Future Vision Future of Computer Vision Horst Bischof Inst. for Computer Graphics and Vision Graz.

Slides:



Advertisements
Similar presentations
Professor Horst Cerjak, Thomas Pock A Duality Based Approach for Realtime TV-L 1 Optical Flow ICG A Duality Based Approach for Realtime TV-L.
Advertisements

Igor Markov Face Detection and Classification on Mobile Devices.
T1.1- Analysis of acceleration opportunities and virtualization requirements in industrial applications Bologna, April 2012 UNIBO.
COMPUTER-AIDED SURGICAL PLANNING AND PROCEDURES A.Schaeffer; PolyDimensions GmbH, Bickenbach.
Learning with Inference for Discrete Graphical Models Nikos Komodakis Pawan Kumar Nikos Paragios Ramin Zabih (presenter)
New idea presentation. To take leading position in new growing technologies of handheld projectors (Pico DLP) and augmented reality. To take position.
Teaching Courses in Scientific Computing 30 September 2010 Roger Bielefeld Director, Advanced Research Computing.
Richard Yu.  Present view of the world that is: Enhanced by computers Mix real and virtual sensory input  Most common AR is visual Mixed reality virtual.
Discrete-Continuous Optimization for Large-scale Structure from Motion David Crandall, Andrew Owens, Noah Snavely, Dan Huttenlocher Presented by: Rahul.
ICG Professor Horst Cerjak, Thomas Pock Variational Methods for 3D Reconstruction Thomas Pock 1, Chrisopher Zach 2 and Horst Bischof 1 1 Institute.
Virtual Dart: An Augmented Reality Game on Mobile Device Supervisor: Professor Michael R. Lyu Prepared by: Lai Chung Sum Siu Ho Tung.
Programming with CUDA, WS09 Waqar Saleem, Jens Müller Programming with CUDA and Parallel Algorithms Waqar Saleem Jens Müller.
Invited Talk Telepresence in the Real World Presented by: Weihong Li --- ACM Multimedia 2004 Conference Workshop on Effective Telepresence (ETP’04) Duffie.
2007Theo Schouten1 Introduction. 2007Theo Schouten2 Human Eye Cones, Rods Reaction time: 0.1 sec (enough for transferring 100 nerve.
© 2004 by Davi GeigerComputer Vision January 2004 L1.1 Introduction.
IPhone 4 The iphone is a mobile phone with a lot of extras, theses include a camera video recording and MP3 player not to forget the mobile web browsing.
Artificial Intelligence
Facial Recognition CSE 391 Kris Lord.
Computer Vision Systems for the Blind and Visually Disabled. STATS 19 SEM Talk 3. Alan Yuille. UCLA. Dept. Statistics and Psychology.
Computer and Internet Basics.
Artificial Intelligence
Research Area B Leif Kobbelt. Communication System Interface Research Area B 2.
Projective Texture Atlas for 3D Photography Jonas Sossai Júnior Luiz Velho IMPA.
TRENDS IN MOBILE DEVICES TAINE MURRAY CANDIDATE: 3592 CENTRE NAME: WILDERN SCHOOL CENTRE NUMBER:
MACHINE VISION GROUP Graphics hardware accelerated panorama builder for mobile phones Miguel Bordallo López*, Jari Hannuksela*, Olli Silvén* and Markku.
“Low-Power, Real-Time Object- Recognition Processors for Mobile Vision Systems”, IEEE Micro Jinwook Oh ; Gyeonghoon Kim ; Injoon Hong ; Junyoung.
--Caesar Cai TEXT RECOGNITION SENIOR CAPSTONE 2012.
By Arun Bhandari Course: HPC Date: 01/28/12. GPU (Graphics Processing Unit) High performance many core processors Only used to accelerate certain parts.
CSCE 5013 Computer Vision Fall 2011 Prof. John Gauch
OpenCL based machine learning labeling of biomedical datasets Oscar Amoros, Sergio Escalera and Anna Puig Computer Vision Center, Universitat Autònoma.
Programming Concepts in GPU Computing Dušan Gajić, University of Niš Programming Concepts in GPU Computing Dušan B. Gajić CIITLab, Dept. of Computer Science.
Accelerating image recognition on mobile devices using GPGPU
Chapter 5 Software Maran Illustrated Computers CIS 102.
Obstacle Avoidance using Machine Vision Joose Rautemaa
80 million tiny images: a large dataset for non-parametric object and scene recognition CS 4763 Multimedia Systems Spring 2008.
GPU Architecture and Programming
Problems in large-scale computer vision David Crandall School of Informatics and Computing Indiana University.
BAGGING ALGORITHM, ONLINE BOOSTING AND VISION Se – Hoon Park.
12/7/10 Looking Back, Moving Forward Computational Photography Derek Hoiem, University of Illinois Photo Credit Lee Cullivan.
Quick starter Which of the emerging techs do you think have the most potential? To make money?
+ Big Data IST210 Class Lecture. + Big Data Summary by EMC Corporation ( More videos that.
Revision Input Process output Software System software Application software Hardware is the physical part of a computer system,including all machines and.
What is Data Communication? Data communication is the process of collecting and distributing data(text, voice, graphics, video, etc) electrically from.
CDVS on mobile GPUs MPEG 112 Warsaw, July Our Challenge CDVS on mobile GPUs  Compute CDVS descriptor from a stream video continuously  Make.
Motion Estimation using Markov Random Fields Hrvoje Bogunović Image Processing Group Faculty of Electrical Engineering and Computing University of Zagreb.
CONTENT FOCUS FOCUS INTRODUCTION INTRODUCTION COMPONENTS COMPONENTS TYPES OF GESTURES TYPES OF GESTURES ADVANTAGES ADVANTAGES CHALLENGES CHALLENGES REFERENCE.
Introduction to Digital Media 1. What is digital media? Digital media is a form of electronic media where data is stored in digital (as opposed to analog)
A Framework for Perceptual Studies in Photorealistic Augmented Reality Martin Knecht 1, Andreas Dünser 2, Christoph Traxler 1, Michael Wimmer 1 and Raphael.
Application development process Part 1. Overview State of the mobile industry Size of the market Popularity of platforms How users use their devices Internationalisation.
1.Computer Fundamentals Parallel and distributed computing, from algorithms, languages, compilers, architectures to systems. 2.Information Systems Software.
A Globally Optimal Algorithm for Robust TV-L 1 Range Image Integration Christopher Zach VRVis Research Center Thomas Pock, Horst Bischof.
Conclusions on CS3014 David Gregg Department of Computer Science
Deep Learning: What is it good for? R. Burgmann
Large-scale Machine Learning
Robust and Adaptive Approaches to Scene and Object Recognition: An example of a successful cooperative Project Horst Bischof Inst. f. Graphics and Vision.
MSC projects for for CMSC5720(term1), CMSC5721(term2)
COMP61011 : Machine Learning Ensemble Models
Detecting Room Occupancy with Pi Camera
Li Fei-Fei, UIUC Rob Fergus, MIT Antonio Torralba, MIT
Pearson Lanka (Pvt) Ltd.
Facial Recognition [Biometric]
DICOM 11/21/2018.
Graphical User Interface Based Digital Sixth Sense
East China Normal University Fang Li
The Pitfalls and Guidelines for Mobile ML Benchmarking
CS4670: Intro to Computer Vision
History of Deep Learning 1/16/19
Machine Learning based Data Analysis
THE ASSISTIVE SYSTEM SHIFALI KUMAR BISHWO GURUNG JAMES CHOU
Computer Science Dr Hwang Chair, Computer Science Department
Presentation transcript:

ICG Professor Horst Cerjak, Horst Bischof Future Vision Future of Computer Vision Horst Bischof Inst. for Computer Graphics and Vision Graz University of Technology

ICG Professor Horst Cerjak, Horst Bischof Future Vision Motto of the talk It is a fantastic time …

ICG Professor Horst Cerjak, Horst Bischof Future Vision Motto of the talk to do computer vision!

ICG Professor Horst Cerjak, Horst Bischof Future Vision WHY?

ICG Professor Horst Cerjak, Horst Bischof Future Vision Computer Vision At least three goals Understand biological visual systems Build machines that see What are the fundamental processes of seeing

ICG Professor Horst Cerjak, Horst Bischof Future Vision Computer Vision The systems today are still exceedingly limited in their performance  considerable room for improvement Where are chairs? Two interpretations? How many feet?

ICG Professor Horst Cerjak, Horst Bischof Future Vision Holy Grails in Vision 1.Segmentation 2. Correspondence Recognition Problem

ICG Professor Horst Cerjak, Horst Bischof Future Vision Future of Computer Vision Where do the innovations come from? 1. Hardware 2. Algorithms/Software

ICG Professor Horst Cerjak, Horst Bischof Future Vision HARDWARE

ICG Professor Horst Cerjak, Horst Bischof Future Vision Hardware First time that HW is no longer a real limitation !! Processing Image Resolution Storage Internet Mobile Devices Networks of cameras

ICG Professor Horst Cerjak, Horst Bischof Future Vision Processing Moore’s Law still holds! Multi-core CPUs Highly Parallel  GPUs (+ Software eg. Cuda) DEMO

ICG Professor Horst Cerjak, Horst Bischof Future Vision Resolution 1900 Chicago & Alton Railroad Train (photograph a train), $5000 Ever growing resolution: 1975: 100 x 100 = 0.01 MP 2008: 9216 × 9216 = 85 MP (BAE) UltraCam x : 216 MP New fantastic opportunities  Computational Cameras

ICG Professor Horst Cerjak, Horst Bischof Future Vision Internet Huge repository of images Flickr: 3.Nov ~3 Billion Photos On-line 1 Million added a day YouTube: new Videos a day 20% of Internet Traffic What can we do with these images?

ICG Professor Horst Cerjak, Horst Bischof Future Vision Mobile Vision Most of us have a mobile CV device with them Small Cameras Embedded Systems Mobile CV next large application area  Place Recognition  Recognizing Tags  Shopping  Games  Augmented Reality 4,4mmx15mm

ICG Professor Horst Cerjak, Horst Bischof Future Vision ALGORITHMS

ICG Professor Horst Cerjak, Horst Bischof Future Vision (Some) New Developments Bayesian MethodsEnergy Minimization Discrete Continuous Machine Learning & Vision

ICG Professor Horst Cerjak, Horst Bischof Future Vision Bayesian Methods Lots of applications Computationally heavy Easily parallelizable  Energy minimization approaches Ill-posed Prior Data

ICG Professor Horst Cerjak, Horst Bischof Future Vision Energy minimization Level SetsConvex formulationsGraph cuts Continuous Discrete Local optimaGlobal optima * GPU ImplementationMemory limitations Metrication errors Apapted f. D Cremers 2007 More to come

ICG Professor Horst Cerjak, Horst Bischof Future Vision Continous energy functional Data term potentially non-convex  Global Optimal Solution Defines domain of application –Denoising –Segmentation –Stereo Total Variation regularization Data term Pock et.al

ICG Professor Horst Cerjak, Horst Bischof Future Vision Vision & Learning Combining Computer Vision with ML  Huge Success We have good/stable features SIFT Boundary fragments If enough data learning works SVM Boosting