Igniting Machine Intelligence with Gravity: Exploiting Newton’s Law based High-Impact Machine Learning Techniques For Efficient Feature Extraction

With rapid advancements in technology, effective human–machine interaction has become increasingly important, making accurate face recognition a critical research area. Traditional face recognition systems predominantly rely on single-modality data, which limits their robustness under real-world conditions. To address these limitations, multimodal face recognition—integrating information from multiple sources such as visual and audio data—has gained significant attention.Despite extensive research, face recognition remains challenging due to variations in illumination, noise, rotation, and occlusion. This thesis addresses these challenges by proposing novel algorithms for invariant feature detection. A key contribution is a new edge detection technique inspired by Newton’s universal law of gravitational force. The method computes gravitational interactions based on signal variation direction and magnitude, and derives vector sums in horizontal and vertical directions to extract precise facial edges.