SpikeSen: Low-Latency In-Sensor-Intelligence Design With Neuromorphic Spiking Neurons

In-sensor-processing (ISP) paradigm has been exploited in state-of-the-art vision system designs to pave the way towards power-efficient sensing and processing. The redundant data transmission between sensors and processors is significantly minimized by local computation within each pixel. However, existing ISP designs suffer from limited frame rates and degraded fill factors. In this brief, we introduce a low-latency in-sensor-intelligence neuromorphic vision system using neuromorphic spiking neurons, namely <monospace>SpikeSen</monospace>. <monospace>SpikeSen</monospace> directly operates on the photocurrents and executes the computation in the frequency domain, reducing the long exposure time and speeding up the computation. Experiments show that <monospace>SpikeSen</monospace> can achieve more than <inline-formula> <tex-math notation="LaTeX">$6.1\times $ </tex-math></inline-formula> computation speedup compared to existing ISP designs with competitive energy consumption per pixel.

• Publication year

  2023

• Venue

  IEEE Transactions on Circuits and Systems - II - Express Briefs

• Publication date

  2023-06-01

• Fields of study

  Computer Science, Engineering

• Identifiers
  DOI 10.1109/TCSII.2023.3235888
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Senputing: An Ultra-Low-Power Always-On Vision Perception Chip Featuring the Deep Fusion of Sensing and Computing

  2021influential reference
• Utilizing Direct Photocurrent Computation and 2D Kernel Scheduling to Improve In-Sensor-Processing Efficiency

  2020influential reference
• Ultrafast machine vision with 2D material neural network image sensors

  2020cited by this paper
• 5.1 A Stacked Global-Shutter CMOS Imager with SC-Type Hybrid-GS Pixel and Self-Knee Point Calibration Single Frame HDR and On-Chip Binarization Algorithm for Smart Vision Applications

  2019cited by this paper
• A Low-Power Convolutional Neural Network Face Recognition Processor and a CIS Integrated With Always-on Face Detector

  2018influential reference
• Loihi: A Neuromorphic Manycore Processor with On-Chip Learning

  2018cited by this paper
• 4.9 A 1ms high-speed vision chip with 3D-stacked 140GOPS column-parallel PEs for spatio-temporal image processing

  2017influential reference
• RedEye: Analog ConvNet Image Sensor Architecture for Continuous Mobile Vision

  2016cited by this paper
• TrueNorth: Design and Tool Flow of a 65 mW 1 Million Neuron Programmable Neurosynaptic Chip

  2015cited by this paper
• Neurogrid: A Mixed-Analog-Digital Multichip System for Large-Scale Neural Simulations

  2014cited by this paper
• 7.2 243.3pJ/pixel bio-inspired time-stamp-based 2D optic flow sensor for artificial compound eyes

  2014influential reference
• Arbitrated Time-to-First Spike CMOS Image Sensor With On-Chip Histogram Equalization

  2007cited by this paper
• A High Dynamic Range CMOS Image Sensor With Inpixel Light-to-Frequency Conversion

  2006cited by this paper
• CMOS Schmitt trigger design

  1994cited by this paper

• APRIS: Approximate Processing ReRAM In-Sensor Architecture Enabling Artificial-Intelligence-Powered Edge

  2025cites this paper
• Advancements in neuromorphic computing for bio-inspired artificial vision: A review

  2025cites this paper
• Designing and Training Neural Networks for Analog In-Sensor Deployment: A Hardware-Aware Analysis

  2025cites this paper
• Exploiting Temporal-Unrolled Parallelism for Energy-Efficient SNN Acceleration

  2024cites this paper