Energy-Efficient VLSI Hardware for Edge AI in Image Processing

There is an increasing need for low-power Very Large-Scale Integration (VLSI) technology as many AI applications are deployed in the network model, all of which require continuous image processing. This research work proposes a comprehensive technique that combines hardware acceleration, low-energy neural network algorithms, better memory hierarchies, and efficient system architecture design. Since it conserves energy without any performance loss, the proposed technique is well-suited for low-powered peripheral devices. The testing results reveal that the proposed VLSI hardware solution outperforms earlier solutions in terms of energy efficiency. Furthermore, it analyzes the problems and tradeoffs inherent in the design process and makes recommendations for further research.