Abhinay Dubasi

Developed a state-of-the-art deep learning model in Python using Tensor Flow and Keras. Implemented a modified version of the EfficientNet-B3 architecture for enhanced feature extraction from dermoscopic images, integrating deep transfer learning techniques to improve model performance. Applied advanced data preprocessing techniques including image augmentation and normalization to augment the dataset for gaining validation accuracy around 90.6%.

Introduced an autonomous street light system that operates based on sensor readings and logic gates. Developed a system to detect objects or motion within a sensor's range, triggering automated light activation. Configured the light to remain on while the object is detected and turn off when out of range. Designed the system using digital electronics principles, incorporating logic gates for efficient operation. Automated the lighting process, eliminating the need for manual control and enhancing energy efficiency.