Efficient CNN Hardware Architecture Based on Linear Approximation and Computation Reuse Technique

Computer Science

• Convolutional Neural Network 100%
• Approximation (Algorithm) 100%
• Hardware Architecture 100%
• Deep Neural Network 40%
• Network Design 40%
• Trained Network 20%
• Sparsity 20%
• Power Consumption 20%
• Hardware Resource 20%
• Research Effort 20%
• Field Programmable Gate Arrays 20%
• Neural Network Model 20%
• Computational Complexity 20%
• Reference Model 20%
• Quantization 20%
• Hardware Accelerator 20%
• Memory Access 20%
• Neural Network Architecture 20%
• Data Reuse 20%
• Resource Utilisation 20%