Recent research:
The researcher develops integrated approaches at the intersection of artificial intelligence, robotics, computer vision, and biomechanical sensing to advance sport performance analysis, line-following robotics, medical image processing, and postural control in the elderly. Across the publications, the work emphasizes autonomous classification and interpretation of human movement and health-related images using inertial sensors, advanced neural networks (including Efficient-UNet and SVM/MLP classifiers), and streamlined hardware-software solutions. The methods address practical challenges in real-world contexts: automatic squat technique classification via multi-sensor inertial capture and machine learning; a durable line-follower robot with a compact PCB-based control system; automated segmentation of vitreous hemorrhages in fundus images using a thresholding pre-processing step followed by an Efficient-UNet architecture; and a systematic review of inertial-sensor-based stabilometric analysis for postural control in the elderly. Collectively, these works contribute to improving accuracy, reliability, and applicability of AI-driven sensing and analysis in sports science, healthcare, and assistive technology, highlighting a trajectory toward deployable, low-cost, and data-driven solutions in rehabilitation, mechanistic understanding of movement, and clinical decision support.