This is the authors' abstract. We don't add key points for this paper.
Monocular depth estimation allows to extract the depth map from a single image. State-of-the-art algorithm MiDaS allows to extract the depth map from a single image, however its results are not accurate for video sequences. We propose a new monocular depth estimation method based on MiDaS. Quality improvement is obtained by utilizing information not only from the current video frame but also from neighboring frames. Alternative ways to extract information from neighboring frames are presented and their quality compared. Presented method works in real-time and produces more accurate predictions than alternative methods in the majority of cases.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Varlamova et al. (2022) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: