An Approach for Hairstyle Recommendation using Machine Learning

Aesthetic assessments indicate that hair is a distinctive feature that can significantly enhance facial appearance. According to beauty experts, approximately 70% of a person's overall facial look is influenced by their haircut or hairstyle. Haircuts play a crucial role in shaping women's self-perception, making the choice of a suitable hairstyle a challenging decision. This paper introduces an innovative framework for selecting the ideal hairstyle or haircut by first classifying the face shape. The model integrates facial shape analysis, expertise in hairstyles, and hair length to recommend the most fitting options. Focused primarily on women's hairstyles, this framework addresses a specific segment within the broader field. Recognizing face shape is considered a vital preliminary step for choosing the appropriate hairstyle. The proposed model classifies face shapes from user-uploaded portraits using machine learning techniques to identify facial landmarks. A Naïve Bayes classifier then recommends suitable hairstyles based on the detected face shape, hair length, and input from hair experts. Users can share their recommended styles with beauticians through "The Beauty Quest" Salon network platform. The system was trained with 5,000 images using Python, achieving a 91% accuracy rate in face shape classification and an 83% accuracy rate in hairstyle recommendations.