Object detection is the problem of two steps. First, it should localize an object or multiple objects inside an image. Secondly, it gives out a predicted class for each of the localized objects. There have been several object detection methods that use a sliding window-based approach. One of the popular detection techniques is face detection approach, developed by Viola and Jones[1]. The paper exploited the fact that the human face has strong descriptive features such as regions near eyes which are darker than near the mouth. So there may be a significant difference between the rectangle area surrounding the eyes with respect to the rectangular area near the nose. Using this as one of the several pre-defined patterns of rectangle pairs, their method computed area difference between rectangles in each pattern.
Detecting faces is a two-step process:
...
