Presentation 2002/6/20
Skin Color Recovery from Over-Exposed Images Using Neural Networks
Stephen Karungaru, Minoru Fukumi, Norio Akamatsu,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we introduce a new method to recover and reconstruct skin color from images over-exposed to scene illumination. Over-exposure is a normal and serious problem in photography especially in images taken under direct sunlight or when extra illumination is added to a scene using a camera's flash. The skin color regions most susceptible to over-exposure include areas around the forehead, tip of the nose and the cheeks. The recovery of skin color can find applications in many areas of computer vision for example face detection, recognition and facial expressions analysis. Our method uses the feed-forward neural network and a distance measure method to learn and construct the skin color face image map. The distance measure method we used is the Cylindrical Metric method (CMM). We learnt the relationships between neighboring pixels skin color distance and their relative position in a face. We then used this information to reconstruct over-exposed skin color regions.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Neural networks / Cylindrical Metric Method / Over-exposure / skin color reconstruction
Paper # NC2002-11
Date of Issue

Conference Information
Committee NC
Conference Date 2002/6/20(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Skin Color Recovery from Over-Exposed Images Using Neural Networks
Sub Title (in English)
Keyword(1) Neural networks
Keyword(2) Cylindrical Metric Method
Keyword(3) Over-exposure
Keyword(4) skin color reconstruction
1st Author's Name Stephen Karungaru
1st Author's Affiliation Department of Information Science and Intelligent Systems, Faculty of Engineering, University of Tokushima()
2nd Author's Name Minoru Fukumi
2nd Author's Affiliation Department of Information Science and Intelligent Systems, Faculty of Engineering, University of Tokushima
3rd Author's Name Norio Akamatsu
3rd Author's Affiliation Department of Information Science and Intelligent Systems, Faculty of Engineering, University of Tokushima
Date 2002/6/20
Paper # NC2002-11
Volume (vol) vol.102
Number (no) 157
Page pp.pp.-
#Pages 6
Date of Issue