Multi-task Video Enhancement for Dental Interventions

TL;DR

This paper presents a deep learning network for multi-task video enhancement in dental procedures, improving visual quality by addressing low-light, noise, and motion issues, and demonstrating state-of-the-art results in near real-time.

Contribution

A novel multi-task deep network for dental video enhancement that jointly performs restoration and temporal alignment, with a new dataset for further research.

Findings

Achieves state-of-the-art results in multiple enhancement tasks

Operates in near real-time for practical dental applications

Introduces the first dental video dataset with multi-task labels

Abstract

A microcamera firmly attached to a dental handpiece allows dentists to continuously monitor the progress of conservative dental procedures. Video enhancement in video-assisted dental interventions alleviates low-light, noise, blur, and camera handshakes that collectively degrade visual comfort. To this end, we introduce a novel deep network for multi-task video enhancement that enables macro-visualization of dental scenes. In particular, the proposed network jointly leverages video restoration and temporal alignment in a multi-scale manner for effective video enhancement. Our experiments on videos of natural teeth in phantom scenes demonstrate that the proposed network achieves state-of-the-art results in multiple tasks with near real-time processing. We release Vident-lab at https://doi.org/10.34808/1jby-ay90, the first dataset of dental videos with multi-task labels to facilitate further research in relevant video processing applications.