A Brief Overview of Computer Vision - PowerPoint PPT Presentation

Loading...

PPT – A Brief Overview of Computer Vision PowerPoint presentation | free to download - id: 69b3ff-OTNhM



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

A Brief Overview of Computer Vision

Description:

A Brief Overview of Computer Vision Jinxiang Chai – PowerPoint PPT presentation

Number of Views:117
Avg rating:3.0/5.0
Slides: 37
Provided by: jchai
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: A Brief Overview of Computer Vision


1
A Brief Overview of Computer Vision
  • Jinxiang Chai

2
What is Computer Vision?
  • Computer vision is the science and technology of
    machines that see.
  • Concerned with the theory for building artificial
    systems that obtain information from images.
  • The image data can take many forms, such as a
    video sequence, views from multiple cameras, or
    multi-dimensional data from a medical scanner

3
Applications
  • Robot perception (e.g. an industrial robot or an
    autonomous vehicle, autonomous helicopter,
    humanoid robots).

4
Honda ASIMO Humanoid Robot
  • Face detection
  • Face recognition
  • Posture/gesture recognition (e.g., hand waving)
  • Environment recognition (e.g., obstacles)

5
Applications
  • Robot perception (e.g. an industrial robot or an
    autonomous vehicle, humanoid robots).
  • Detecting events (e.g. for visual surveillance or
    people counting).

6
Detecting Events
  • Customer tracking and activity analysis

7
Applications
  • Robot perception (e.g. an industrial robot or an
    autonomous vehicle, humanoid robots).
  • Detecting events (e.g. for visual surveillance or
    people counting).
  • Modeling objects or environments

8
Modeling objects or environments
  • Modeling buildings, plants, faces, cars etc.

9
Applications
  • Robot perception (e.g. an industrial robot or an
    autonomous vehicle, humanoid robots).
  • Detecting events (e.g. for visual surveillance or
    people counting).
  • Modeling objects or environments
  • Interaction (e.g. as the input to a device for
    computer-human interaction).

10
Interactions
  • Interactions with computers and video games, etc.

Face recognition for automatic login
Computer vision for game interfaces (Sony eyetoy,
Microsoft Kinect)
11
Applications
  • Robot perception (e.g. an industrial robot or an
    autonomous vehicle, humanoid robots).
  • Detecting events (e.g. for visual surveillance or
    people counting).
  • Modeling objects or environments
  • Interaction (e.g. as the input to a device for
    computer-human interaction).
  • Organizing information (e.g. for indexing
    databases of images and image sequences).

12
Organizing information
  • Flickr (www. Flickr.com) has 3 billion images
  • Youtube has tons of videos.
  • Need new ways to search, analyze, summarize a
    large collection of internet images and videos

13
Image Representation
  • An image is a 2D rectilinear array of Pixels
  • - A width X height array where each entry of
    the array stores a single pixel

14
Image Representation
  • A pixel stores color information
  • Luminance pixels
  • - gray-scale images (intensity images)
  • - 0-255
  • - 8 bits per pixel
  • Red, green, blue pixels (RGB)
  • - Color images
  • - Each channel 0-255
  • - 24 bits per pixel

15
Image Representation
  • An image is a 2D rectilinear array of Pixels
  • - A width X height array where each entry of
    the array stores a single pixel
  • - Each pixel stores color information

(255,255,255)
16
Images
  • Which kind of information you can obtain from
    images

17
Images
  • Which kind of information you can obtain from
    images

Edge detection
18
Images
  • Which kind of information you can obtain from
    images

Edge detection
Corner feature detection
19
Images
  • Which kind of information you can obtain from
    images

Edge detection
Corner feature detection
Geometric primitive detection
20
Images
  • Which kind of information you can obtain from
    images

Edge detection
Corner feature detection
Geometric primitive detection
Object detection
21
Images
  • Which kind of information you can obtain from
    images

Edge detection
Corner feature detection
Geometric primitive detection

Face alignment and recognition
Object detection
22
How about multiple images?
  • What can we obtain if we have multiple images?

23
How about multiple images?
  • What can we obtain if we have multiple images?

Two images of the same scene
24
Structure and motion analysis
  • Given two or more images of the same scene or
    object, estimate camera motion and 3D object
    structure (e.g., depth)

unknown camera viewpoints
25
Structure and motion analysis
  • Given two or more images of the same scene or
    object, estimate camera motion and 3D object
    structure (e.g., depth)

unknown camera viewpoints
How to estimate camera parameters? - where is
the camera? - where is it pointing? - what
are internal parameters, e.g. focal length?
26
Structure and motion analysis
  • Given two or more images of the same scene or
    object, estimate camera motion and 3D object
    structure (e.g., depth)

unknown camera viewpoints
How to estimate camera parameters? - where is
the camera? - where is it pointing? - what
are internal parameters, e.g. focal length?
Camera calibration!
27
Structure and motion analysis
  • Reconstruct the depth information.

Input images
How to find the depth information of this point?
28
Structure and motion analysis
  • Reconstruct the depth information.

Input images
How to find the depth information of this point?
- find the corresponding point in the right
image.
29
Structure and motion analysis
  • Reconstruct the depth information.

Input images
30
Structure and motion analysis
  • Reconstruct the depth information.

Input images
Depth images
31
Structure and motion analysis
  • Reconstruct 3D models from multiple images

Reconstruction results from 23 images
32
All together video
  • Click here
  • - feature detection
  • - feature matching (epipolar geometry)
  • - structure from motion
  • - stereo reconstruction
  • - triangulation
  • - texture mapping

33
How about video sequences?
  • What can we obtain from video?

34
How about video sequences?
  • What can we obtain from video?

Optical flow where are pixels moving to?
35
How about multiple video sequences
  • Modeling dynamic objects (video click here)

36
Modeling human motion from video
  • Single-view camera
  • Interactively construct human motion form video
About PowerShow.com