Reconfigurable Optical Dilated Convolution Accelerator for Multi‐Scale Feature Extraction

ABSTRACT Dilated convolution is pivotal for dense prediction tasks such as semantic segmentation, as it expands the receptive field to capture multi‐scale contextual information without increasing parameter complexity. However, its electronic implementation often faces bottlenecks due to non‐contiguous memory access and high computational load. Here, we report a Reconfigurable Optical Dilated Convolution Accelerator (RODCA) based on a wavelength‐dispersion co‐tuning mechanism. By dynamically controlling optical frequency comb spacing and dispersion, RODCA enables reconfigurable dilation rates while improving spectral efficiency. The system achieves an equivalent computing throughput of 13.858 Tera operations per second (TOPS). We experimentally validate the RODCA using Laplacian kernels and semantic segmentation tasks, where it achieves a pixel accuracy of 88.46% and a mean Intersection over Union (mIoU) of 76.08%, closely matching in‐silico results. This work demonstrates RODCA's capability for efficient multi‐scale feature extraction, highlighting its potential for next‐generation photonic computing in complex AI applications.

Authors

Institutions

Publication Details

Journal
Laser & Photonics Review
Published
2026-09-15
DOI
https://doi.org/10.1002/lpor.71892
Primary Topic
Neural Networks and Reservoir Computing
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Reconfigurable Optical Dilated Convolution Accelerator for Multi‐Scale Feature Extraction

Arnan Mitchell, Roberto Morandotti, Xingyuan Xu, Yunping Bai et al.
Laser & Photonics Review
Neural Networks and Reservoir Computing
article

Reconfigurable Optical Dilated Convolution Accelerator for Multi‐Scale Feature Extraction

Arnan Mitchell, Roberto Morandotti, Xingyuan Xu, Yunping Bai, Kun Xu, Yixuan Zheng, Sai T. Chu, Sha Zhu, Xiaotian Zhu, Shifan Chen, Brent E. Little, David Moss, Xin Peng, Tian Zhang, Yifu Xu, Y. Zheng, Zhihui Liu, Ruilin Liao, Xiyao Jia, Yuyang Liu, Yue Zhou
article en

Abstract

ABSTRACT Dilated convolution is pivotal for dense prediction tasks such as semantic segmentation, as it expands the receptive field to capture multi‐scale contextual information without increasing parameter complexity. However, its electronic implementation often faces bottlenecks due to non‐contiguous memory access and high computational load. Here, we report a Reconfigurable Optical Dilated Convolution Accelerator (RODCA) based on a wavelength‐dispersion co‐tuning mechanism. By dynamically controlling optical frequency comb spacing and dispersion, RODCA enables reconfigurable dilation rates while improving spectral efficiency. The system achieves an equivalent computing throughput of 13.858 Tera operations per second (TOPS). We experimentally validate the RODCA using Laplacian kernels and semantic segmentation tasks, where it achieves a pixel accuracy of 88.46% and a mean Intersection over Union (mIoU) of 76.08%, closely matching in‐silico results. This work demonstrates RODCA's capability for efficient multi‐scale feature extraction, highlighting its potential for next‐generation photonic computing in complex AI applications.

Laser & Photonics Review
Beijing University of Posts and Telecommunications (CN), City University of Hong Kong (HK), Nankai University (CN), Institut National de la Recherche Scientifique (CA), Yangtze Optical Electronic (China) (CN), Applied Photonics (United Kingdom) (GB), Medical Technologies (Czechia) (CZ), Swinburne University of Technology (AU), RMIT University (AU), Technical University of Denmark (DK)
Openalex Percentile: Top 8%
Neural Networks and Reservoir Computing
DOISourceOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.