Deep Learning-Based RGB-Thermal Image Denoising: Review and Applications

Research output: Journal PublicationArticlepeer-review


Recently, vision-based detection (VD) technology has been well-developed, and its general-purpose object detection algorithms have been applied in various scenes. VD can be divided into two categories based on the type of modality: single-modal (single RGB or single thermal) and bimodal. Image denoising is typically the first stage of image processing in VD, where redundant information and noisy data are removed to produce clearer images for effective object detection. This study reviews deep learning-based image denoising for RGB and thermal images, investigating the denoising procedure, methodologies, and performances of algorithms tested with benchmark datasets. After introducing denoising models, the main results on public RGB and thermal datasets are presented and analyzed, and conclusions of objective comparison in practical effect are drawn. This review can serve as a reference for researchers in RGB–infrared denoising, image restoration, and related fields.
Original languageEnglish
Number of pages29
JournalMultimedia Tools and Applications
Publication statusPublished - 29 Jun 2023


  • Image denoising
  • Deep learning
  • Computer vision
  • Object detection
  • Thermal imaging


Dive into the research topics of 'Deep Learning-Based RGB-Thermal Image Denoising: Review and Applications'. Together they form a unique fingerprint.

Cite this