-
L-Sort: An Efficient Hardware for Real-time Multi-channel Spike Sorting with Localization
Authors:
Yuntao Han,
Shiwei Wang,
Alister Hamilton
Abstract:
Spike sorting is essential for extracting neuronal information from neural signals and understanding brain function. With the advent of high-density microelectrode arrays (HDMEAs), the challenges and opportunities in multi-channel spike sorting have intensified. Real-time spike sorting is particularly crucial for closed-loop brain computer interface (BCI) applications, demanding efficient hardware…
▽ More
Spike sorting is essential for extracting neuronal information from neural signals and understanding brain function. With the advent of high-density microelectrode arrays (HDMEAs), the challenges and opportunities in multi-channel spike sorting have intensified. Real-time spike sorting is particularly crucial for closed-loop brain computer interface (BCI) applications, demanding efficient hardware implementations. This paper introduces L-Sort, an hardware design for real-time multi-channel spike sorting. Leveraging spike localization techniques, L-Sort achieves efficient spike detection and clustering without the need to store raw signals during detection. By incorporating median thresholding and geometric features, L-Sort demonstrates promising results in terms of accuracy and hardware efficiency. We assessed the detection and clustering accuracy of our design with publicly available datasets recorded using high-density neural probes (Neuropixel). We implemented our design on an FPGA and compared the results with state of the art. Results show that our designs consume less hardware resource comparing with other FPGA-based spike sorting hardware.
△ Less
Submitted 26 June, 2024;
originally announced June 2024.
-
Towards Highly Expressive Machine Learning Models of Non-Melanoma Skin Cancer
Authors:
Simon M. Thomas,
James G. Lefevre,
Glenn Baxter,
Nicholas A. Hamilton
Abstract:
Pathologists have a rich vocabulary with which they can describe all the nuances of cellular morphology. In their world, there is a natural pairing of images and words. Recent advances demonstrate that machine learning models can now be trained to learn high-quality image features and represent them as discrete units of information. This enables natural language, which is also discrete, to be join…
▽ More
Pathologists have a rich vocabulary with which they can describe all the nuances of cellular morphology. In their world, there is a natural pairing of images and words. Recent advances demonstrate that machine learning models can now be trained to learn high-quality image features and represent them as discrete units of information. This enables natural language, which is also discrete, to be jointly modelled alongside the imaging, resulting in a description of the contents of the imaging. Here we present experiments in applying discrete modelling techniques to the problem domain of non-melanoma skin cancer, specifically, histological images of Intraepidermal Carcinoma (IEC). Implementing a VQ-GAN model to reconstruct high-resolution (256x256) images of IEC images, we trained a sequence-to-sequence transformer to generate natural language descriptions using pathologist terminology. Combined with the idea of interactive concept vectors available by using continuous generative methods, we demonstrate an additional angle of interpretability. The result is a promising means of working towards highly expressive machine learning systems which are not only useful as predictive/classification tools, but also means to further our scientific understanding of disease.
△ Less
Submitted 9 July, 2022;
originally announced July 2022.
-
An FPGA Implementation of Convolutional Spiking Neural Networks for Radioisotope Identification
Authors:
Xiaoyu Huang,
Edward Jones,
Siru Zhang,
Shouyu Xie,
Steve Furber,
Yannis Goulermas,
Edward Marsden,
Ian Baistow,
Srinjoy Mitra,
Alister Hamilton
Abstract:
This paper details the FPGA implementation methodology for Convolutional Spiking Neural Networks (CSNN) and applies this methodology to low-power radioisotope identification using high-resolution data. Power consumption of 75 mW has been achieved on an FPGA implementation of a CSNN, with an inference accuracy of 90.62% on a synthetic dataset. The chip validation method is presented. Prototy** wa…
▽ More
This paper details the FPGA implementation methodology for Convolutional Spiking Neural Networks (CSNN) and applies this methodology to low-power radioisotope identification using high-resolution data. Power consumption of 75 mW has been achieved on an FPGA implementation of a CSNN, with an inference accuracy of 90.62% on a synthetic dataset. The chip validation method is presented. Prototy** was accelerated by evaluating SNN parameters using SpiNNaker neuromorphic platform.
△ Less
Submitted 24 February, 2021;
originally announced February 2021.
-
Spiking Neural Network Based Low-Power Radioisotope Identification using FPGA
Authors:
Xiaoyu Huang,
Edward Jones,
Siru Zhang,
Shouyu Xie,
Steve Furber,
Yannis Goulermas,
Edward Marsden,
Ian Baistow,
Srinjoy Mitra,
Alister Hamilton
Abstract:
this paper presents a detailed methodology of a Spiking Neural Network (SNN) based low-power design for radioisotope identification. A low power cost of 72 mW has been achieved on FPGA with the inference accuracy of 100% at 10 cm test distance and 97% at 25 cm. The design verification and chip validation methods are presented. It also discusses SNN simulation on SpiNNaker for rapid prototy** and…
▽ More
this paper presents a detailed methodology of a Spiking Neural Network (SNN) based low-power design for radioisotope identification. A low power cost of 72 mW has been achieved on FPGA with the inference accuracy of 100% at 10 cm test distance and 97% at 25 cm. The design verification and chip validation methods are presented. It also discusses SNN simulation on SpiNNaker for rapid prototy** and various considerations specific to the application such as test distance, integration time, and SNN hyperparameter selections.
△ Less
Submitted 25 October, 2020;
originally announced October 2020.
-
Event-based Signal Processing for Radioisotope Identification
Authors:
Xiaoyu Huang,
Edward Jones,
Siru Zhang,
Steve Furber,
Yannis Goulermas,
Edward Marsden,
Ian Baistow,
Srinjoy Mitra,
Alister Hamilton
Abstract:
This paper identifies the problem of unnecessary high power overhead of the conventional frame-based radioisotope identification process and proposes an event-based signal processing process to address the problem established. It also presents the design flow of the neuromorphic processor.
This paper identifies the problem of unnecessary high power overhead of the conventional frame-based radioisotope identification process and proposes an event-based signal processing process to address the problem established. It also presents the design flow of the neuromorphic processor.
△ Less
Submitted 2 October, 2020; v1 submitted 11 July, 2020;
originally announced July 2020.
