Compiling CNNs with Cain: focal-plane processing for robot navigation
| Autor: | Edward Stow, Abrar Ahsan, Yingying Li, Ali Babaei, Riku Murai, Sajad Saeedi, Paul H. J. Kelly |
|---|---|
| Přispěvatelé: | Engineering & Physical Science Research Council (E, Engineering & Physical Science Research Council (EPSRC) |
| Jazyk: | angličtina |
| Rok vydání: | 2022 |
| Předmět: |
Technology
Science & Technology 1702 Cognitive Sciences Analogue computing Robotics Computer Science Artificial Intelligence Convolution SIMD Industrial Engineering & Automation Artificial Intelligence Computer Science Edge inference 0801 Artificial Intelligence and Image Processing Image sensor 0913 Mechanical Engineering |
| Popis: | Focal-plane Sensor-processors (FPSPs) are a camera technology that enables low power, high frame rate computation in the image sensor itself, making them suitable for edge computation. To fit into the sensor array, FPSPs are highly resource-constrained, with limited instruction set and few registers - which makes developing complex algorithms difficult. In this work, we present Cain, a compiler for convolutional filters that targets SCAMP-5, a general-purpose FPSP. Cain generates code to evaluate multiple convolutional kernels at the same time. It generates code that avoids the need for hardware multipliers, while orchestrating the exploitation of common sub-terms—leading to a large reduction in instruction count compared to both straightforward and prior optimized approaches. We demonstrate the capability enabled by Cain on SCAMP-5 with robotic navigation for near-sensor high-speed and low-power computation, by using Cain to implement a neural network on the focal plane. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|