Technique | Contour shape descriptor |
---|---|
Document | MPEG-7 Visual (ISO/IEC 15938-3) |
Name | Miroslaw Bober (Mitsubishi Electric ITE) |
miroslaw.bober@vil.ite.mee.com | |
Contact | miroslaw.bober@vil.ite.mee.com |
Type | Application |
External Libraries | none |
Related Ds/DSs | Still region DS |
Used Ds/DSs | none |
Input | Binary shape masks |
Extraction | Yes |
Client Appl | Search & Retrieval |
Summary | The Contour-based shape descriptor describes shape features of a closed contour. It uses the curvature scale space representation and eccentricity and circularity of the original and filtered shape. |
Strong Points | The descriptor is rotational, scale and position invariant. It is robust to affine shape deformations and deformations due to non-rigid motion (both very common in video sequences) The descriptor has good shape generalisation properties, similar to human perception of visual shapes. It is very compact (110 bits on average for the MPEG-7 shape database). |
Limitations | Here are some limitations of the current implementation: Deafult white shape on black background (can be changed) If more than one contour is in the image, the descripotr is extracted from the first contour found in the image (upper-left, top left). |
Known Problems | None |
Parameters | Uses the general XM-SW parameters. |