Table of Contents
Partial Shape Similarity
0. Motivation
1. Edge Detection
2. Edge Grouping
3. Extraction of Contour Parts
4. Recognizing Shapes by Parts
5. Results
Structural Shape Similarity
0. Introduction
1. Shape Segmentation
2. Skeleton Computation
3. Skeleton Pruning
4. Recognizing Shapes by Structures
Contour Grouping Based on Local Symetry
0. Introduction
1. From Edges To Contours
2. Center points
3. Reference Shape Model
4. Particle Filter
5.1. Experimental Result 1
5.2. Experimental Result 2
5.3. Experimental Result 3
5.4. Experimental Result 4
5.5. Experimental Result 5
5.6. Experimental Result 6
5.7. Experimental Result 7
5.8. Experimental Result 8
6. Conclusion
7. Contour Grouping Video
Original image Edge Detection Edge Detection

Edge Detection



Canny Edge Detector
A Canny Edge Detector is a program used to detect object edges in images based on gradient differences. It outputs a grayscale bitmap wherein gradient differences are represented by white lines. These lines approximate edges of objects in an image. This serves us well here as it provides us with a usable contour for objects in images.

[Read More On The Canny Edge Detector]

Edge detection is a process wherein the edges of objects in digital images are isolated as being edges. The purpose of edge detection is to provide a contour for comparison that is as free from extra data, such as color data, as possible. In our project we use a Canny edge detector to isolate edges in an image.

Before: A typical image from the InternetAfter: The image after it has been processed by a Canny Edge Detector
Before and After: An original image (left) and the same
image after it has gone through edge detection (right).

Click each for a larger view.

The Canny edge detector outputs a grayscale bitmap image file, where detected edges are represented by gray and white pixels. The edge detector colors edges on the grayscale based on edge strength, or the clarity with which an edge can be detected. The bitmap file output by the edge detector is, in its simplest form, a matrix of 8-bit gray values. Because of this matrix nature, these bitmaps can easily be processed by a machine. A list of the positions of edge pixels represents a machine-processable, pixel-by-pixel edge contour of a digital image.