Car Recognition Through SIFT Keypoint Matching

1
Car Recognition Through SIFT Keypoint Matching
Third Quarter Report
Drew Stebbins, Pd. 6
2
Background

• License plate tracking technology
• Need for civilian vehicle recognition
• Object detection techniques
• SIFT keypoint matching

3
Problem statement

• Car recognition
• fast moving
• multi-coloured
• complex shapes
• must identify in real-time

4
Region of Interest Identification

• Hough circle transform
• multiple passes

5
(No Transcript)
6
SIFT keypoint matching

• Keypoint detection
• scale-space extrema detection
• keypoint localization
• orientation assignment
• keypoint descriptor

7
(No Transcript)
8
SIFT keypoint matching

• Matching to database images
• nearest-neighbour identification based on
  keypoint description

9
Expectations

• side profile car recognition and categorization
• invariant to lighting changes, also small amounts
  of rotation, translation
• multiple cars
• moderate detection speed

10
Progress

• Region of interest identification
• Hough transform circle detector
• Scale space extrema detection
• Gaussian smoothing function and
  difference-of-Gaussian

11
Plan for Fourth Quarter

• Finish implementation of SIFT keypoint detection
  algorithm
• Test recognition accuracy on database of testing
  images taken from TJ student parking lot

12
You just won the game!