Abstract: Extracting and tracking faces in image sequences is a required first step in many applications, such as facial recognition, facial expression classification, and face tracking. It is a challenging problem in the field of computer vision because of many factors that affect the image, including luminosity, different face colors, background patterns, face orientation, and variability in size, shape, and expression. The objective of this paper is to test a wide range of parameters for a HOG face detector, to establish the most suitable kernel for a Support Vector Machine (SVM). The reliability of this method is then compared with well-known methods for face detection. The aim of this study is to evaluate the performance of the HOG descriptor for detecting a face, experimenting with different kernels to determine the tuned parameters for HOG descriptors for detecting a face. It is found that the HOG in conjunction with SVM scores the highest values for precision, accuracy, and sensitivity: 0.8824, 0.9986, and 0.75 respectively, compared to the Viola-Jones method, which scores 0.6512, 0.9973, and 0.7, and the skin-color method, which scores 0.3968, 0.9947 and 0.625.