本篇論文以抗光影變化的物件偵測來偵測移動人物。此方法結合抗光影變化的物件偵測技術
中所用到的色彩和紋理特徵。其中利用紋理資訊對光照強度不敏感的特性可以去除陰影等光影變
化，運用色彩資訊特徵可以加強前景分離的準確度。因此，可以此物件偵測的結果
, 執行人物偵測。

A novel illumination-invariant human detection algorithm is introduced in this paper. It is based on the moving object detection. The object detection algorithm can be robust and invariant to light intensity by
means of both color and texture features. Owing to the fact that the texture features are not sensitive to the variation of light, shadow effect can then be removed. Moreover, color features is employed to enhance the accuracy of separating foreground from background. Based on the proposed object detection algorithm, human detection can be easily performed and lead to satisfactory results.