Deep Learning-Based Women Hairstyle Classification Through Inception Resnet-101

Artificial intelligence and computer science's computer vision field is revolutionizing a number of industries, including healthcare, automotive, agriculture, security, and entertainment, by enabling robots to assess visual input. For tasks like object detection, classification, and 3D reconstruction, it makes use of methods from the fields of image processing, machine learning, deep learning, and pattern recognition. One promising application of this technology is the classification of hairstyles using pre-trained networks, which can be used for the detection of criminals who may be using certain hairstyles as a form of disguise or it can be used to identify a person through CCTV surveillance for further investigations. In this paper, a study using the performance of Inception Resnet 101, a popular learning architecture has been used for the classification of several common hairstyles. These hairstyles include ‘Bald’, ‘Braids’, ‘Free Hair’, ‘Pony Tail’ and ‘Short Hair’. Validation accuracy of about 89.7% and testing accuracy of 95.6% has been achieved. This approach demonstrates that the model can attain high accuracy even in the face of difficult real-world circumstances, such as changes in illumination, angles, and distances. A few of the industries that need to take this into account are marketing, trend analysis, and customized haircuts. Approaches utilizing computer vision and artificial intelligence facilitate innovation, enhance decision-making, and enhance the user experience.