(†: equal contribution, ~: corresponding author)

📜 Conference Paper

• YOLOv8-LSD: Improved YOLOv8 Focused on Small Target Information Extraction for Road Damage Detection [paper]
  Tao Lin, Qingwang Wang~, Jiangbo Huang, Xin Qu, Gao Ju, Hua Wu
  2024 2nd International Conference on Pattern Recognition, Machine Vision and Intelligent Algorithms (PRMVIA).
  Changsha, China. May, 2024.

Under submission

• FASR-Net: Unsupervised Shadow Removal Leveraging Inherent Frequency Priors [paper] [code]
  Tao Lin, Qingwang Wang~, Qiwei Liang, Minghua Tang, Yuxuan Sun
  

  FASR-Net, an innovative unsupervised network for shadow removal that leverages the frequency characteristics of shadow regions. Key features include a Wavelet Attention Downsampling Module (WADM) for enhanced shadow detail and novel loss functions—frequency loss, brightness-chromaticity loss, and alignment loss—to boost performance. Our experiments on the AISTD and SRD datasets demonstrate that FASR-Net surpasses many existing unsupervised and supervised methods.
  

💡 Ongoing Project

• The idea is currently being conceived and will be updated soon.