OPEN CV – AN INTRODUCTION TO THE VISION LIBRARY AND ITS APPLICATIONS Hemanth Haridas Co founder  VidPulp Technologies
WHAT IS THIS TALK ABOUT? An basic introduction to Open CV, a cross platform vision library. To get students interested in Vision and related fields. Introduction to the capabilities of Open CV. Applications of the library. Limitations of the library
WHAT IS THIS TALK NOT ABOUT? Not a programmatic tutorial on Open CV. Not a detailed explaination about Open CV. No math behind Open CV covered.
 
COMPUTER VISION Rapidly growing field because of cheaper and more capable cameras and affordable processing power. Vision algorithms are starting to mature. Open CV has helped computer vision grow as a field Open CV helps jump-start research by providing them with a computer vision and machine learning infrastructure
CONTENTS Introduction Who Uses OpenCV? What is Computer Vision? How complex is the problem? Details about the Library. Summary
INTRODUCTION OpenCV  ( Open  Source  C omputer  V ision)  library of programming functions for real time computer vision.  (computer efficiency) Written in optimized C and C++ Runs in windows, Linux and MAC OS. Can develop in C, C++, python, ruby, matlab
INTRODUCTION Simple to use infra  Helps to build fairly complicated applications. 500 functions spanning Factory product inspection  medical imaging Security user interface  camera calibration stereo vision robotics contains a full, general-purpose Machine Learning Library (MLL)
WHO USES OPEN CV? Surveillance images and video on the Web. (Flickr, picasa  face recognition and tagging) aerial and street-map images (such as in Google’s Street View) make heavy use of camera calibration and image stitching techniques safety monitoring, unmanned flying vehicles, or biomedical analysis. manufacturing: virtually everything that is mass-produced has been automatically  inspected at some point using computer vision.
WHO USES OPEN CV? license for OpenCV has been structured such that you can build a commercial product using all or part of OpenCV. You are under no obligation to opensource large user community that includes people from major companies (IBM, Microsoft , Intel, SONY, Siemens, and Google, to name only a few) and research centers (such as Stanford, MIT, CMU, Cambridge, and INRIA).
WHO USES OPEN CV? http://groups.yahoo.com/group/OpenCV   -  20,000 members OpenCV was a key part of the vision system in the robot from Stanford, “Stanley”, which won the $2M DARPA Grand Challenge desert robot race web maps, image scan alignment, medical image noise reduction, object analysis, security and intrusion detection systems, automatic monitoring and safety systems, manufacturing inspection systems, camera calibration, military applications, and unmanned aerial, ground, and underwater vehicles
WHAT IS COMPUTER VISION? transformation of data from a still or video camera into either a decision or a new representation. Data - > “the camera is mounted in a car” or “laser range fi nder indicates an object is 1 meter away”. Decision -> “there is a person in this scene” or “there are 14 tumor cells on this slide” new representation -> turning a color image into a grayscale image
HOW HARD CAN THAT BE? Human  brain divides the vision signals to many channels. Identifies important parts Complex feedback mechanism that is little understood. draw on cross-associations made from years of living in the world. Controls lighting through the iris.
HOW A MACHINE SEES IT? A 2d Image of a 3d object. No definite way to reconstruct the 3d image.
HOW A MACHINE SEES IT? Images are corrupted by noise and distortions. (weather, lighting, reflections, movements) Additional contextual knowledge is used  Which is helpful in matching
NOISE Edge Detection -> impossible to detect edges by comparing a point to its neighbours If the comparison is made over a localized area its easier. Compensating noise by using statistics over time. Explicit models learnt from available data. (lens distortions)
CONTEXTUAL INFORMATION The decision taken by vision algorithms depend on the application it is used for. Security system that alerts if a person tries to cross a fence. monitoring system that counts how many people cross through an area in an amusement park. Strategy for vision algos in security cameras different from that of in robots. The more constrained our context , the better the solution will be.
ABOUT THE LIBRARY aimed at providing the basic tools needed to solve computer vision problems high-level functionalities in the library will be sufficient to solve the more complex problems in computer vision the basic components in the library are complete enough to enable creation of a complete solution After you develop a first draft solution, check for weakness and fix it.
OPEN CV TIMELINE
COMPONENTS OF OPENCV
RESOURCES Download Open CV -  http://sourceforge.net/projects/opencvlibrary/ Install guide and tutorials -  http://opencv.willowgarage.com/wiki/ IDE – eclipse, .net , VC++
SUMMARY Open CV is a open source library to implement Computer vision algorithms Computer vision is a complex problem. Made easy with enough context information. Computer vision is a interesting and a fast growing field and skills in this field is niche , in demand
OCR ALGORITHMS Optical character recognition, to recognize text in scanned documents. Useful in detecting text in videos For extracting contextual information in videos. Tesseract and GOCR – open source OCRs available
SPEECH TO TEXT ALGOS Algorithms to convert from Speech to text. Text to speech conversion algos are available. Language translation research is underway.
QUESTIONS?

Open Cv – An Introduction To The Vision

  • 1.
    OPEN CV –AN INTRODUCTION TO THE VISION LIBRARY AND ITS APPLICATIONS Hemanth Haridas Co founder VidPulp Technologies
  • 2.
    WHAT IS THISTALK ABOUT? An basic introduction to Open CV, a cross platform vision library. To get students interested in Vision and related fields. Introduction to the capabilities of Open CV. Applications of the library. Limitations of the library
  • 3.
    WHAT IS THISTALK NOT ABOUT? Not a programmatic tutorial on Open CV. Not a detailed explaination about Open CV. No math behind Open CV covered.
  • 4.
  • 5.
    COMPUTER VISION Rapidlygrowing field because of cheaper and more capable cameras and affordable processing power. Vision algorithms are starting to mature. Open CV has helped computer vision grow as a field Open CV helps jump-start research by providing them with a computer vision and machine learning infrastructure
  • 6.
    CONTENTS Introduction WhoUses OpenCV? What is Computer Vision? How complex is the problem? Details about the Library. Summary
  • 7.
    INTRODUCTION OpenCV ( Open Source C omputer V ision) library of programming functions for real time computer vision. (computer efficiency) Written in optimized C and C++ Runs in windows, Linux and MAC OS. Can develop in C, C++, python, ruby, matlab
  • 8.
    INTRODUCTION Simple touse infra Helps to build fairly complicated applications. 500 functions spanning Factory product inspection medical imaging Security user interface camera calibration stereo vision robotics contains a full, general-purpose Machine Learning Library (MLL)
  • 9.
    WHO USES OPENCV? Surveillance images and video on the Web. (Flickr, picasa face recognition and tagging) aerial and street-map images (such as in Google’s Street View) make heavy use of camera calibration and image stitching techniques safety monitoring, unmanned flying vehicles, or biomedical analysis. manufacturing: virtually everything that is mass-produced has been automatically inspected at some point using computer vision.
  • 10.
    WHO USES OPENCV? license for OpenCV has been structured such that you can build a commercial product using all or part of OpenCV. You are under no obligation to opensource large user community that includes people from major companies (IBM, Microsoft , Intel, SONY, Siemens, and Google, to name only a few) and research centers (such as Stanford, MIT, CMU, Cambridge, and INRIA).
  • 11.
    WHO USES OPENCV? http://groups.yahoo.com/group/OpenCV - 20,000 members OpenCV was a key part of the vision system in the robot from Stanford, “Stanley”, which won the $2M DARPA Grand Challenge desert robot race web maps, image scan alignment, medical image noise reduction, object analysis, security and intrusion detection systems, automatic monitoring and safety systems, manufacturing inspection systems, camera calibration, military applications, and unmanned aerial, ground, and underwater vehicles
  • 12.
    WHAT IS COMPUTERVISION? transformation of data from a still or video camera into either a decision or a new representation. Data - > “the camera is mounted in a car” or “laser range fi nder indicates an object is 1 meter away”. Decision -> “there is a person in this scene” or “there are 14 tumor cells on this slide” new representation -> turning a color image into a grayscale image
  • 13.
    HOW HARD CANTHAT BE? Human brain divides the vision signals to many channels. Identifies important parts Complex feedback mechanism that is little understood. draw on cross-associations made from years of living in the world. Controls lighting through the iris.
  • 14.
    HOW A MACHINESEES IT? A 2d Image of a 3d object. No definite way to reconstruct the 3d image.
  • 15.
    HOW A MACHINESEES IT? Images are corrupted by noise and distortions. (weather, lighting, reflections, movements) Additional contextual knowledge is used Which is helpful in matching
  • 16.
    NOISE Edge Detection-> impossible to detect edges by comparing a point to its neighbours If the comparison is made over a localized area its easier. Compensating noise by using statistics over time. Explicit models learnt from available data. (lens distortions)
  • 17.
    CONTEXTUAL INFORMATION Thedecision taken by vision algorithms depend on the application it is used for. Security system that alerts if a person tries to cross a fence. monitoring system that counts how many people cross through an area in an amusement park. Strategy for vision algos in security cameras different from that of in robots. The more constrained our context , the better the solution will be.
  • 18.
    ABOUT THE LIBRARYaimed at providing the basic tools needed to solve computer vision problems high-level functionalities in the library will be sufficient to solve the more complex problems in computer vision the basic components in the library are complete enough to enable creation of a complete solution After you develop a first draft solution, check for weakness and fix it.
  • 19.
  • 20.
  • 21.
    RESOURCES Download OpenCV - http://sourceforge.net/projects/opencvlibrary/ Install guide and tutorials - http://opencv.willowgarage.com/wiki/ IDE – eclipse, .net , VC++
  • 22.
    SUMMARY Open CVis a open source library to implement Computer vision algorithms Computer vision is a complex problem. Made easy with enough context information. Computer vision is a interesting and a fast growing field and skills in this field is niche , in demand
  • 23.
    OCR ALGORITHMS Opticalcharacter recognition, to recognize text in scanned documents. Useful in detecting text in videos For extracting contextual information in videos. Tesseract and GOCR – open source OCRs available
  • 24.
    SPEECH TO TEXTALGOS Algorithms to convert from Speech to text. Text to speech conversion algos are available. Language translation research is underway.
  • 25.