A dedicated bit-serial hardware neuron for massively-parallel neural networks in fast epilepsy diagnosis

摘要

This paper outlines the feasibility of detecting epilepsy though low-cost and low-energy dedicated hardware with bit-serial processing. The concept of a novel bit-serial data processing unit (DPU) is presented which implements the functionality of a complete neuron. The proposed approach has been tested using various network configurations and compared with related work. The proposed DPU uses only 24 Adaptive Logic Modules on an Altera Cyclone V FPGA. An array of these DPUs are controlled by a simple finite state machine. The proposed DPU allows the construction of complex hardware ANNs that can be implemented in portable equipment that suits the needs of a single epileptic patient in his or her daily activities to detect impending seizure events.