Significance of Conventional and Nonconventional Features in Classification of Face through Human Intelligence
Vani Hiremani, Kishore Kumar Senapati · 2019 International Conference on Smart Systems and Inventive Technology (ICSSIT) · 2019
Feature extraction is the most vital stage in edifice of any system under image processing. The prominent features play important role in the process of image recognition. Features are broadly classified as Conventional and Non-Conventional. The type of features is being selected affects the image classification phase. The state of the art work has achieved commendable success under face recognition using conventional features like skin tone, face shape, eyes, eyebrows, nose and mouth etc.,. Moving slight away from traditional approach of working with only conventional facial features the underlying work aimed to establish an empirical research to understand the significant role of non-conventional features like dressing style, physic, moustache, style of applying vermillion, regional amulet thread, traditional jewellery along with conventional features in face recognition process. These two features can effectively be used to pursue emotions, behaviour and geographic regional information of a person who is under surveillance. Human intellect is significantly effective in recognising face even in unfavourable conditions like occlusion, blurred image and variation in pose. In this research work ample empirical evidences are gathered and are analysed quantitatively and qualitatively to comprehend that not only the conventional features but many non-conventional features prompted human intelligence to make out native information of a person in given face image. Human Intelligence ensured 96% accuracy in this research work.