Context-based object viewpoint estimation: A 2D relational approach

The task of object viewpoint estimation has been a challenge since the early days of computer vision. To estimate the viewpoint (or 2pose) of an object, people have mostly looked at object intrinsic features, such as shape or appearance. Surprisingly, informative rfeatures provided by other, extrinsic elements in the scene, have so far mostly been ignored. At the same time, contextual cues have pbeen proven to be of great benefit for related tasks such as object detection or action recognition. In this paper, we explore how Ainformation from other objects in the scene can be exploited for viewpoint estimation. In particular, we look at object configurations 1 by following a relational neighbor-based approach for reasoning about object relations. We show that, starting from noisy object 2detections and viewpoint estimates, exploiting the estimated viewpoint and location of other objects in the scene can lead to improved object viewpoint predictions. Experiments on the KITTI dataset demonstrate that object configurations can indeed be ]V used as a complementary cue to appearance-based viewpoint estimation. Our analysis reveals that the proposed context-based C method can improve object viewpoint estimation by reducing specific types of viewpoint estimation errors commonly made by . methods that only consider local information. Moreover, considering contextual information produces superior performance in sscenes where a high number of object instances occur. Finally, our results suggest that, following a cautious relational neighbor c[ formulation brings improvements over its aggressive counterpart for the task of object viewpoint estimation.

Jaar van publicatie:2017

BOF-keylabel:ja

IOF-keylabel:ja

BOF-publication weight:1

CSS-citation score:1

Authors from:Higher Education

Toegankelijkheid:Open