Introduction
We adopted the methods from face detection and applied it
for car rear-view detection. We believe that face detection and
car detection are similar in nature and applying face detection
methods for car detection is a natural choice. The success of
face detection may have been due to the fact that face images
have rich internal features. Some image categories, such as pedestrian,
can only be characterized by the contour of the object
since there is not a characteristic pattern that stands out inside
the region of the object. On the other hand, faces have very distinctive
pattern, caused by eyes, nose, mouth, etc, that can be
well captured by simple filters encoding the intensity difference
as in Viola and Jones [1]. Similarly, rear views of cars also have
distinctive patterns, such as the dark shadow region right below
the car, and dark tire region, as shown in section 3.1. Therefore,
we decided to apply the methods developed for face detection to
car detection.
IntroductionWe adopted the methods from face detection and applied itfor car rear-view detection. We believe that face detection andcar detection are similar in nature and applying face detectionmethods for car detection is a natural choice. The success offace detection may have been due to the fact that face imageshave rich internal features. Some image categories, such as pedestrian,can only be characterized by the contour of the objectsince there is not a characteristic pattern that stands out insidethe region of the object. On the other hand, faces have very distinctivepattern, caused by eyes, nose, mouth, etc, that can bewell captured by simple filters encoding the intensity differenceas in Viola and Jones [1]. Similarly, rear views of cars also havedistinctive patterns, such as the dark shadow region right belowthe car, and dark tire region, as shown in section 3.1. Therefore,we decided to apply the methods developed for face detection tocar detection.
การแปล กรุณารอสักครู่..
