Spiking neural networks fine-tuning for brain image segmentation
IntroductionThe field of machine learning has undergone a significant transformation with the progress of deep artificial neural networks (ANNs) and the growing accessibility of annotated data. ANNs usually require substantial power and memory usage to achieve optimal performance. Spiking neural net...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-11-01
|
| Series: | Frontiers in Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnins.2023.1267639/full |
| _version_ | 1797643304065564672 |
|---|---|
| author | Ye Yue Marc Baltes Nidal Abuhajar Tao Sun Avinash Karanth Charles D. Smith Trevor Bihl Jundong Liu |
| author_facet | Ye Yue Marc Baltes Nidal Abuhajar Tao Sun Avinash Karanth Charles D. Smith Trevor Bihl Jundong Liu |
| author_sort | Ye Yue |
| collection | DOAJ |
| description | IntroductionThe field of machine learning has undergone a significant transformation with the progress of deep artificial neural networks (ANNs) and the growing accessibility of annotated data. ANNs usually require substantial power and memory usage to achieve optimal performance. Spiking neural networks (SNNs) have recently emerged as a low-power alternative to ANNs due to their sparsity nature. Despite their energy efficiency, SNNs are generally more difficult to be trained than ANNs.MethodsIn this study, we propose a novel three-stage SNN training scheme designed specifically for segmenting human hippocampi from magnetic resonance images. Our training pipeline starts with optimizing an ANN to its maximum capacity, then employs a quick ANN-SNN conversion to initialize the corresponding spiking network. This is followed by spike-based backpropagation to fine-tune the converted SNN. In order to understand the reason behind performance decline in the converted SNNs, we conduct a set of experiments to investigate the output scaling issue. Furthermore, we explore the impact of binary and ternary representations in SNN networks and conduct an empirical evaluation of their performance through image classification and segmentation tasks.Results and discussionBy employing our hybrid training scheme, we observe significant advantages over both ANN-SNN conversion and direct SNN training solutions in terms of segmentation accuracy and training efficiency. Experimental results demonstrate the effectiveness of our model in achieving our design goals. |
| first_indexed | 2024-03-11T14:12:47Z |
| format | Article |
| id | doaj.art-037345e9eac2414e834c1fa45e68eecb |
| institution | Directory Open Access Journal |
| issn | 1662-453X |
| language | English |
| last_indexed | 2024-03-11T14:12:47Z |
| publishDate | 2023-11-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroscience |
| spelling | doaj.art-037345e9eac2414e834c1fa45e68eecb2023-11-01T16:21:45ZengFrontiers Media S.A.Frontiers in Neuroscience1662-453X2023-11-011710.3389/fnins.2023.12676391267639Spiking neural networks fine-tuning for brain image segmentationYe Yue0Marc Baltes1Nidal Abuhajar2Tao Sun3Avinash Karanth4Charles D. Smith5Trevor Bihl6Jundong Liu7School of Electrical Engineering and Computer Science, Ohio University, Athens, OH, United StatesSchool of Electrical Engineering and Computer Science, Ohio University, Athens, OH, United StatesSchool of Electrical Engineering and Computer Science, Ohio University, Athens, OH, United StatesCentrum Wiskunde and Informatica (CWI), Machine Learning Group, Amsterdam, NetherlandsSchool of Electrical Engineering and Computer Science, Ohio University, Athens, OH, United StatesDepartment of Neurology, University of Kentucky, Lexington, KY, United StatesDepartment of Biomedical, Industrial and Human Factors Engineering, Wright State University, Dayton, OH, United StatesSchool of Electrical Engineering and Computer Science, Ohio University, Athens, OH, United StatesIntroductionThe field of machine learning has undergone a significant transformation with the progress of deep artificial neural networks (ANNs) and the growing accessibility of annotated data. ANNs usually require substantial power and memory usage to achieve optimal performance. Spiking neural networks (SNNs) have recently emerged as a low-power alternative to ANNs due to their sparsity nature. Despite their energy efficiency, SNNs are generally more difficult to be trained than ANNs.MethodsIn this study, we propose a novel three-stage SNN training scheme designed specifically for segmenting human hippocampi from magnetic resonance images. Our training pipeline starts with optimizing an ANN to its maximum capacity, then employs a quick ANN-SNN conversion to initialize the corresponding spiking network. This is followed by spike-based backpropagation to fine-tune the converted SNN. In order to understand the reason behind performance decline in the converted SNNs, we conduct a set of experiments to investigate the output scaling issue. Furthermore, we explore the impact of binary and ternary representations in SNN networks and conduct an empirical evaluation of their performance through image classification and segmentation tasks.Results and discussionBy employing our hybrid training scheme, we observe significant advantages over both ANN-SNN conversion and direct SNN training solutions in terms of segmentation accuracy and training efficiency. Experimental results demonstrate the effectiveness of our model in achieving our design goals.https://www.frontiersin.org/articles/10.3389/fnins.2023.1267639/fullspiking neural network (SNN)ANN-SNN conversionimage segmentationfine-tuningU-Net |
| spellingShingle | Ye Yue Marc Baltes Nidal Abuhajar Tao Sun Avinash Karanth Charles D. Smith Trevor Bihl Jundong Liu Spiking neural networks fine-tuning for brain image segmentation Frontiers in Neuroscience spiking neural network (SNN) ANN-SNN conversion image segmentation fine-tuning U-Net |
| title | Spiking neural networks fine-tuning for brain image segmentation |
| title_full | Spiking neural networks fine-tuning for brain image segmentation |
| title_fullStr | Spiking neural networks fine-tuning for brain image segmentation |
| title_full_unstemmed | Spiking neural networks fine-tuning for brain image segmentation |
| title_short | Spiking neural networks fine-tuning for brain image segmentation |
| title_sort | spiking neural networks fine tuning for brain image segmentation |
| topic | spiking neural network (SNN) ANN-SNN conversion image segmentation fine-tuning U-Net |
| url | https://www.frontiersin.org/articles/10.3389/fnins.2023.1267639/full |
| work_keys_str_mv | AT yeyue spikingneuralnetworksfinetuningforbrainimagesegmentation AT marcbaltes spikingneuralnetworksfinetuningforbrainimagesegmentation AT nidalabuhajar spikingneuralnetworksfinetuningforbrainimagesegmentation AT taosun spikingneuralnetworksfinetuningforbrainimagesegmentation AT avinashkaranth spikingneuralnetworksfinetuningforbrainimagesegmentation AT charlesdsmith spikingneuralnetworksfinetuningforbrainimagesegmentation AT trevorbihl spikingneuralnetworksfinetuningforbrainimagesegmentation AT jundongliu spikingneuralnetworksfinetuningforbrainimagesegmentation |