Gender and Age Detection Techniques for Blind People using Principal Component Analysis

Predicting someone's gender and age-based solely on appearance is challenging for an AI model, as it requires recognizing and understanding complex patterns in images and videos. This kind of prediction is particularly used for blind people. Different kinds of techniques are applied to analyze and predict gender and age classification with the help of images. In the previous methods, accuracy and live prediction have different challenges such as accuracy and prediction rates. This work proposes a hybrid model to predict and analyze gender and age classification. This work consists of Haar Cascade, Histogram of Oriented Gradients, Hessian Filter and Principal Component Analysis. The Haar Cascade algorithm finds the face from the video images with constant speed and time. The Histogram-orientation extracts the features from the images, and the Hessian Filter finds the wrinkles and, based on that, finds the age. Principal Component Analysis is used to visualize and find the patterns from the images. This proposed hybrid work is implemented using 10137 images, and based on that, training and testing are performed. The proposed work is evaluated using recall, f1-score and precision. The proposed work achieved 87.30%. and 86.9% precision for gender classification and age classification respectively. This, compared to the previous work, produced effective results for gender and age detection.