Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 39384 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Kasaie, Shohreh
- Abstract:
- In recent years, there has been a growing research interest in segmenting skin regions in color images. Skin segmentation aims at locating skin regions in an unconstrained input image. Skin detection is considered as an important preprocess in many applications such as face detection, face tracking, and filtering of objectionable web images. The most attractive properties of skin detection include low computational cost, increase of the total processing speed, and being invariance against rotation, scale, partial occlusion, and pose change. Because of the diversity of various image processing tasks, there exists no optimum method that can perform properly for all applications. Most of the existing skin detection approaches are based on skin color. Some challenges of skin detection such as complex background and non-human skin-like objects in image lead to use region-based skin detection methods. In this research, a general structure for region-based skin detection is used. Based on this structure, two new methods are proposed. A region is detected as skin if its color and texture are similar to skin. One of the methods uses the cellular learning automata as an analyzer for skin texture. The proposed algorithm has demonstrated a true positive rate of about 83.4% and a false positive rate of about 11.3% on the Compaq skin database. The other method uses the coefficients of the contourlet transform as the features of skin texture. The proposed algorithm has been tested on provided database that consist of 300 hand-labeled images and achieved 87.7% of true positive rate and 5.3% of false positive rate. The proposed methods are capable of detecting any kind of skin color
- Keywords:
- Image Texture Analysis ; Color Space ; Contourlet Transform ; Stochastic Cellular Learning Automata ; Skin Detection ; Pixel Based Skin Detection ; Texture Based Skin Detection
Digital Object List-
محتواي پايان نامه - view
Bookmark