Event Detail
Unlocking the Third Dimension: a Journey into Depth Estimation from Image
Unlocking the Third Dimension: a Journey into Depth Estimation from Image
Modeling the surrounding environment in its three dimensions with computer vision allows for tackling complex tasks such as navigation and interaction with other agents. The first step toward 3D sensing requires recovery of the third dimension that is missing in the images: depth. This quest has been open for decades, during which several techniques voted for this purpose have been developed, any of them powered by deep learning nowadays. In this talk, we are leaving on a journey in the realm of depth estimation, exploring the different approaches, some of the most representative deep models for each, and the challenges remaining open to the community.