Object Boundary Tracking in Moving Palm Image for Personal Identification

Palm includes enough features for participating the personal identification, one of the encouraging matters of relying on palm features is their resistivity to the age and environmental impacts. In this project, palm based personal identification is implemented based on IIT Delhi palm dataset. Right hand image is being acquired from each candidate using a mobile camera placed inside a black color box. Region of interest was accurately cropped from each palm image and hence deep learning and machine learning algorithms are used for palm classification. Results shown that Feed Forward Neural Network based palm recognition model is outperformed over the other machine learning technologies.