Vident-lab: a dataset for multi-task video processing of phantom dental scenes

We introduce a new, asymmetrically annotated dataset of natural teeth in phantom scenes for multi-task video processing: restoration, teeth segmentation, and inter-frame homography estimation. Pairs of frames were acquired with a beam splitter. The dataset constitutes a low-quality frame, its high-quality counterpart, a teeth segmentation mask, and an inter-frame homography matrix. The homography warps the current frame to the previous frame with respect to the teeth. Moreover, we provide the list of human-annotated segmentation masks so that future segmentation methods can adhere to our evaluation protocol and compare their results to MOST-Net in [1]. The remaining segmentation masks were obtained with HRNet48. The dataset has the training, validation, and test sets of 300, 29, and 80 videos, respectively.

本项研究提出了一组新型、非对称标注的自然牙齿在幻影场景下的多任务视频处理数据集，包括牙齿修复、牙齿分割以及帧间单应性估计。帧对通过分束器获取。该数据集包括低质量帧、其高质量对应帧、牙齿分割掩码以及帧间单应性矩阵。单应性变换将当前帧相对于牙齿扭曲至前帧。此外，我们还提供了人工标注的分割掩码列表，以便未来的分割方法能够遵循我们的评估协议，并将他们的结果与文献[1]中的MOST-Net进行比较。剩余的分割掩码由HRNet48获取。该数据集包含300个训练视频、29个验证视频和80个测试视频。