UltraFast-LiNET: Lightweight network for real-time low-light image enhancement

Researchers have developed UltraFast-LiNET, a highly efficient convolutional neural network designed for real-time low-light image enhancement on resource-constrained edge devices. The network features a novel Dynamic Shift Convolution (DSConv) operation and a Multi-Scale Shift Residual Block (MSRB) to achieve a large receptive field with minimal parameters. To address training challenges in lightweight models, a multi-level gradient-aware loss function was introduced. UltraFast-LiNET demonstrates competitive performance, achieving a PSNR of 19.81 dB on the LOL dataset with as few as 180 learnable parameters, balancing real-time processing with image quality for edge deployment. AI

IMPACT This research offers a new lightweight model for real-time image enhancement on edge devices, potentially improving performance in applications like autonomous driving and mobile photography.

RANK_REASON This is a research paper detailing a new model and methodology for image enhancement. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• DSConv
• Dynamic Shift Convolution
• LOL dataset
• Multi-Scale Shift Residual Block
• UltraFast-LiNET
• Yuhan Chen

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →