An improvement of mouth detection implementing color features extraction method for facial images with lighting problems

Color features extraction is one among methods widely used for mouth extraction. This relates to specific features possessed by mouth color compared to others facial components. Mouth has a strong chromatic color for red but weak for blue. Color features extraction produces colors that are used to generate a mouth map. This map is required for mouth extraction. Some previous researches on mouth extraction are not produce maximal results yet. Mouth components that were extracted are not complete due to lighting effect captured on images. This study aimed at providing improvement of mouth detection especially for image with lighting effect. The improvement is conducted through the implementation of different correction factors on the existing mouth map formula. Result shows that implementation of the proposed method achieve an accuracy of detection 95%.