FPGA-Accelerated Real-Time Image Reconstruction For CMOS Electrochemical Sensing Arrays
Year: 2026
Type: conference-paper
Abstract: This paper presents a real-time, high-speed decompression and image reconstruction platform for Ion-Sensitive Field-Effect Transistor (ISFET) arrays, implemented on FPGA. The system leverages compressed sensing (CS) principles and employs an Iterative Hard Thresholding (IHT) algorithm to reconstruct ISFET images from highly compressed measurements in real time. The hardware implementation reconstructs a 32×32 ISFET image from a compressed signal with a compression ratio of 4:1, achieving a peak signal-to-noise ratio (PSNR) of up to 27.49 dB. The reconstruction latency from the compressed signal input buffer to the output image buffer is 870 μs. The FPGA-based platform integrates a real-time readout and control interface for the ISFET array, along with a user-friendly visualisation interface that displays the reconstructed images dynamically. The platform constitutes a portable and energy-efficient solution for the ISFET System-on-Chip (SoC) enhanced with compressed sensing. It enables configurable image processing parameters to accommodate a range of image sensing applications.
Authors: Tom Zhou, Shuanghua Liu, Pantelis Georgiou
Institution: Imperial College London
Language: English
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7165159366
