Palmistry with Deep Learning Approaches Using CNN and OpenCV

Palmistry, an ancient tradition interpreting hand features, offers insights into personality traits, life potential, and destiny. While automated palmistry solutions exist, they often overlook important factors like gender and hand dominance, resulting in incomplete analyses. This paper proposes a new approach incorporating Convolutional Neural Networks (CNN) for gender and hand dominance detection, along with OpenCV for line identification. The methodology consists of three primary segments: gender identification, hand dominance determination, and palm line detection. For gender detection, a CNN was constructed using TensorFlow and Keras, featuring three convolutionallayers, max-pooling layers, a flatten layer, and two dense layers. This model achieved 99.91% accuracy on the training dataset and 98.24% on the testing dataset. Hand dominance detection utilized a comparable CNN architecture, achieved 100% training accuracy and 99.81% testing accuracy. Palm line detection used OpenCV, involving preprocessing steps such as resizing, grayscale conversion, and smoothing with bilateral and Gaussian filters. The Heart, Head, and Life lines were identified using the Canny edge detector and Hough transform. By bridging the gap between current technology and traditional palmistry, this holistic approach offers a valuable tool for introspection and guidance. By merging traditional practices with advanced computational methods, the approach ensures more precise and comprehensive palmistry analysis, addressing the shortcomings of existing automated solutions.