基於模糊推論之膚色補償方法應用於彩色影像
No Thumbnail Available
Date
2004
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
彩色影像的色彩資訊應用在物體的檢測（例如利用膚色做人臉檢測之
前置處理）是相當有用的訊息，但是應用色彩作為物體檢測的特徵，首先
必須面對的是光線所造成的影響。在過亮、過暗、陰影或偏光的環境下皆
會使物體檢測的檢測率大為降低。
在本文中採用分散式模糊邏輯推論，利用影像之平均亮度與像素之
RGB值，建構分散式模糊推論引擎，分別推論出像素RGB之補償量，而得
到一種適應不同狀況(光線偏暗或偏亮)之光線補償演算法，解決人臉影像
在不均勻光源的影響，並且配合利用模糊關係與模糊統計試驗法所建構之
膚色相似模型，設計人的膚色檢測器，找出人臉膚色的區域。
並且利用嘴唇的顏色特徵，設計一個唇色檢測器與眼球其低灰度值的
特徵，設計眼睛檢測器，配合人臉五官的幾何關係找出人臉的位置，此方
法可減少複雜背景的影響且可解決多人臉重疊在一個區域的情況。
經由實驗得知此模糊推論模式能有效解決彩色影像之光線補償的問
題，且本文所提之演算法亦可處理複雜背景的影像，因而可提高人臉之檢
測率。
Different environment illumination has a great impact on object detection and recognition. The automatic radiation correction of a highlighed object area or lowlighed object area plays an important role in the field of image processing and computer vision. Skin color can be used for human face detection. In the paper, we propose a light compensation method for skin color segmentation under varying light conditions. A major problem of using skin color is that a face region may not be detected under poor or lighting conditions. We adopted a fuzzy logic technique to determine the compensation value the brightness. The exposure control system proposed in the paper uses “RGB” and Y of pixels in the color image determines a compensation amount by the fuzzy reasoning. Based on a light compensation technique and skin color similar model (SSM), skin regions can be detected and then face candidates can be obtained. The characteristics of eyes and lips of faces are used to help to detect each face candidates. The performance of the system is evaluated through assessment experiments. Experimental results show that this proposed methodcan improve the performance of face segmentation under poor or strong light conditions and detect faces with wide variations in size, scale, color, position and expression in images.
Different environment illumination has a great impact on object detection and recognition. The automatic radiation correction of a highlighed object area or lowlighed object area plays an important role in the field of image processing and computer vision. Skin color can be used for human face detection. In the paper, we propose a light compensation method for skin color segmentation under varying light conditions. A major problem of using skin color is that a face region may not be detected under poor or lighting conditions. We adopted a fuzzy logic technique to determine the compensation value the brightness. The exposure control system proposed in the paper uses “RGB” and Y of pixels in the color image determines a compensation amount by the fuzzy reasoning. Based on a light compensation technique and skin color similar model (SSM), skin regions can be detected and then face candidates can be obtained. The characteristics of eyes and lips of faces are used to help to detect each face candidates. The performance of the system is evaluated through assessment experiments. Experimental results show that this proposed methodcan improve the performance of face segmentation under poor or strong light conditions and detect faces with wide variations in size, scale, color, position and expression in images.
Description
Keywords
光線補償, 模糊邏輯, 色彩空間, 膚色相似度模型, 唇色檢測, light compensation, fuzzy logic, color space, skin color similar model, lip detector