Hardware-aware Graph Neural Networks prunning for embedded event-based vision
Quick Answer
This study introduces a hardware-aware pruning and quantization strategy for Graph Convolutional Neural Networks, optimizing them for embedded FPGA platforms.
Quick Take
This study introduces a hardware-aware pruning and quantization strategy for Graph Convolutional Neural Networks, optimizing them for embedded FPGA platforms. The approach achieved BRAM memory reductions of 28.8% for CIFAR-10 with a 1.65% accuracy drop, 31.4% for MNIST-DVS with a 3.55% accuracy decrease, and 26.5% for N-Caltech101 with a 5.18% accuracy reduction, addressing the needs of real-time data processing in mobile robotics.
Key Points
- Proposed method optimizes Graph Convolutional Neural Networks for embedded FPGA platforms.
- Achieved 28.8% BRAM reduction for CIFAR-10 with a 1.65% accuracy decrease.
- Reduced BRAM by 31.4% for MNIST-DVS, with a 3.55% drop in accuracy.
- N-Caltech101 saw a 26.5% BRAM reduction, with a 5.18% accuracy reduction.
- Method addresses low latency and power consumption in dynamic environments.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Event-based cameras are gaining popularity as the sensor of choice for mobile robotics, due to their high performance in dynamic environments. However, these applications require efficient real-time data processing with low latency and power consumption. One strategy to meet these stringent requirements is hardware acceleration of efficient algorithms that preserve the temporal sparsity of event data. In this work, we propose an optimization strategy for Graph Convolutional Neural Networks models aimed at adapting their architecture to the limited resources of embedded heterogeneous FPGA platforms. Our method incorporates hardware-aware pruning and quantization, taking into account the trade-off between on-chip memory savings and inference accuracy. Strategic exploration of the design space with Fine Grid Search and Greedy layer-wise Iterative Deepening Search methods enables flexible adaptation of the model architecture to the target platform. Our approach was evaluated across various network configurations and multiple datasets, resulting in BRAM memory reductions of 28.8% for CIFAR-10 (with a 1.65% decrease in accuracy), 31.4% for MNIST-DVS (accuracy drop of 3.55%), and 26.5% for N-Caltech101 (with a 5.18% accuracy reduction).
| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2607.06739 [cs.CV] |
| (or arXiv:2607.06739v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2607.06739 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | 2025 Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA), Poznan, Poland, 2025, pp. 182-187 |
| Related DOI: | https://doi.org/10.23919/SPA65537.2025.11215154
DOI(s) linking to related resources |
Submission history
From: Tomasz Kryjak [view email]
[v1]
Tue, 7 Jul 2026 19:09:03 UTC (6,070 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.
