Ternary-Input Binary-Weight CNN Accelerator Design for Miniature Object Classification System with Query-Driven Spatial DVS
By: Yuyang Li , Swasthik Muloor , Jack Laudati and more
Potential Business Impact:
Lets tiny cameras see more with less power.
Miniature imaging systems are essential for space-constrained applications but are limited by memory and power constraints. While machine learning can reduce data size by extracting key features, its high energy demands often exceed the capacity of small batteries. This paper presents a CNN hardware accelerator optimized for object classification in miniature imaging systems. It processes data from a spatial Dynamic Vision Sensor (DVS), reconfigurable to a temporal DVS via pixel sharing, minimizing sensor area. By using ternary DVS outputs and a ternary-input, binary-weight neural network, the design reduces computation and memory needs. Fabricated in 28 nm CMOS, the accelerator cuts data size by 81% and MAC operations by 27%. It achieves 440 ms inference time at just 1.6 mW power consumption, improving the Figure-of-Merit (FoM) by 7.3x over prior CNN accelerators for miniature systems.
Similar Papers
RISC-V Based TinyML Accelerator for Depthwise Separable Convolutions in Edge AI
Hardware Architecture
Makes smart devices run faster and use less power.
Flexible Vector Integration in Embedded RISC-V SoCs for End to End CNN Inference Acceleration
Distributed, Parallel, and Cluster Computing
Makes smart devices run AI faster and use less power.
High Clockrate Free-space Optical In-Memory Computing
Emerging Technologies
Makes computers see and learn super fast.