A Survey on Object Detection and Recognition for Blurred and Low-Quality Images: Handling, Deblurring, and Reconstruction

作者：Aetesam Ali Khan Ashar, Ajwad Abrar, Jiangjiang Liu · 年份：2024 · DOI：10.1145/3686397.3686413 · 被引用次数：8 · 研究领域：Advanced Neural Network Applications、Advanced Image and Video Retrieval Techniques、Advanced Image Processing Techniques

Object detection and recognition in blurred and low-quality images present significant challenges in computer vision and image processing. This survey paper provides a comprehensive overview of the state-of-the-art techniques, methodologies, and advancements in addressing these challenges. Blurred and low-quality images are encountered in various real-world scenarios, such as surveillance, medical imaging, and autonomous vehicles, making robust object detection and recognition essential. This paper reviews the key issues, datasets, evaluation metrics, and recent advancements in this field, with an emphasis on deep learning-based approaches. Particular emphasis is placed on the integration of popular object detection models such as SSD (Single Shot Multibox Detector), COCO (Common Objects in Context), and YOLO (You Only Look Once) with deblurring techniques. Through this survey, we aim to provide researchers and practitioners with valuable insights into the current landscape of object detection and recognition in challenging imaging conditions, facilitating further research and application development in this critical domain.