On-Sensor Binarized Fully Convolutional Neural Network with A Pixel Processor Array

TL;DR

This paper introduces a novel on-sensor binarized FCN implementation on a Pixel Processor Array, enabling real-time object segmentation and localization with high speed and low resource usage.

Contribution

It presents the first implementation of an FCN on a PPA sensor, using binarized networks for efficient, embedded, pixel-level processing.

Findings

Achieved over 280 FPS inference speed.

Demonstrated effective coarse segmentation and object localization.

Implemented three convolution layers entirely in pixel processors.

Abstract

This work presents a method to implement fully convolutional neural networks (FCNs) on Pixel Processor Array (PPA) sensors, and demonstrates coarse segmentation and object localisation tasks. We design and train binarized FCN for both binary weights and activations using batchnorm, group convolution, and learnable threshold for binarization, producing networks small enough to be embedded on the focal plane of the PPA, with limited local memory resources, and using parallel elementary add/subtract, shifting, and bit operations only. We demonstrate the first implementation of an FCN on a PPA device, performing three convolution layers entirely in the pixel-level processors. We use this architecture to demonstrate inference generating heat maps for object segmentation and localisation at over 280 FPS using the SCAMP-5 PPA vision chip.