Face Recognition and Face Detection Using Open Computer Vision Classifiers and Python

3. Face Detection Using Haar-Cascades

A mathematical function that produces square-shaped waves with a beginning and an end and used to create box shaped patterns to recognise signals with sudden transformations is called Haar wavelet as shown in an example in figure 1. A cascade can be created by combining several wavelets and that cascade can easily and efficiently identify edges, lines and circles with different colour intensities. These sets are used in Viola Jones face detection technique in 2001 and since then more patterns are introduced for object detection as shown in figure 1. To analyse an image using Haar cascades, target image is identified and a scale is selected smaller than the that image. The scale is then placed on the image. Then the average of the values of pixels in each section of the image is taken. A match is considered if the difference between two values is above a given threshold value. Face detection on a human face is performed similarly by matching a combination of different Haar like features. For example, forehead, eyebrows and eyes contrast as well as the nose with eyes as shown below in figure 2. Accurate face detection cannot be achieved by a single classifier, various different classifiers are combined to get an accurate result as shown in the block diagram in figure 3.

Figure 1.

A Haar wavelet followed by resulting Haar like features