Edge Detection

1
Edge Detection

• Exploring Edge Detection Techniques
• By Asa Kusuma
• Computer Systems Research 2007-2008

2
Basic Edge Detection

• Calculates difference between adjacent pixels
• Averages Difference in the X and Y direction

3
Is This Part of the Image an Edge?
4
No, it's only Noise
5
Analysis of Basic Edge Detection

• Advantages
• Fast
• Easy to code

• Disadvantages
• Only works well with simple images
• Confused by noise
• Narrow view

6
Tackling the Problem of Noise

• What is Noise?
• Random changes in pixel value
• Usually associated with texture
• Like sandpaper
• Detecting Noise
• Looking at the entire image
• Searching for line patterns
• Blurring

7
Blurring Technique

• Blur the image
• Look for differences across larger pixel gaps

8
Recursive Blurring Technique

• Blur image
• Find biggest pixel difference gaps
• Recursively go through and find adjacent gaps

9
Problems and Issues

• Speed
• Blurring takes a lot of time
• Finding lines takes time
• Language Choice
• Currently using Python
• Easy to code different algorithims
• Slow
• C/C?
• Faster

10
Project Direction

• Analysis of running biomechanics
• Pronation
• Image from behind runner
• Angle of ankle mass
• Compare to angle of leg
• Heel Strike
• Image from the side of runner
• Measure bend in knees after impact