Hence, an ANN or a Fuzzy Logic based system will be able to choose the possible make-up colors that suit a particular person, when her image is fed in to the system.
Phase 1: Research on a suitable method for classification of make-up colors, methods of feature extraction (color extraction) from a photograph and facial recognition, how Neural Networking/Fuzzy Logic can be applied to find a solution to the problem and determine the best suited neural network/fuzzy logic approach (e.g. Fuzzy Logic, Supervised Training, un-supervised Training etc) to address the problem. Conduct a literature review on the above areas.
Phase 2: Learn the basics of make-up and feature evaluation with the help of a beautician. Collect a set of sample data for training and testing (this can be obtained from the beautician; the sample test data may include images of a number of different faces and the suitable make-up colors decided by the beautician to match those faces)
Phase 3: Propose a suitable method to extract the color/complexion from a given image of a face, implement the method and test the accuracy. The image captured through a camera may need enhancements to preserve the original skin color. Design and implement the neural network/fuzzy logic based solution to choose make-up colors.
Phase 4: Test the system with a set of suitable test data set. Test data can be prepared with the help of a beautician. Also get the beautician to review the system and get feedback on the level of accuracy.
a) Functional Objectives:
Develop a fairly accurate Neural Network/Fuzzy Logic based system to automate the role of a beautician in deciding make-up colors for a person (precisely a lady) by examining facial features such as skin color, eye color, lip shape etc. This system is intended to help beauticians, as well as any ordinary person in selecting appropriate make-up colors that suits a particular person.
b) Learning Objectives View More »