3D Viewpoint Estimation Based on Aesthetics
نویسندگان
چکیده
منابع مشابه
Towards Viewpoint Invariant 3D Human Pose Estimation
We propose a viewpoint invariant model for 3D human pose estimation from a single depth image. To achieve this, our discriminative model embeds local regions into a learned viewpoint invariant feature space. Formulated as a multi-task learning problem, our model is able to selectively predict partial poses in the presence of noise and occlusion. Our approach leverages a convolutional and recurr...
متن کاملSay Cheese: Personal Photography Layout Recommendation Using 3D Aesthetics Estimation
Many people fail to take exquisite pictures in a beautiful scenery for the lack of professional photography knowledge. In this paper, we focus on how to aid people to master daily life photography using a computational layout recommendation method. Given a selected scene, we first generate several synthetic photos with different layouts using 3D estimation. Then we employ a 3D layout aesthetic ...
متن کاملRobust Wide Baseline Scene Alignment Based on 3D Viewpoint Normalization
This paper presents a novel scheme for automatically aligning two widely separated 3D scenes via the use of viewpoint invariant features. The key idea of the proposed method is following. First, a number of dominant planes are extracted in the SfM 3D point cloud using a novel method integrating RANSAC and MDL to describe the underlying 3D geometry in urban settings. With respect to the extracte...
متن کاملReal-time Free-viewpoint Video Based on 3D Shape Reconstruction
In this paper, we present a system generating free-viewpoint video in real-time using multiple cameras and a PC-cluster. Our system firstly reconstructs a shape model of objects by the visual cone intersection method, secondly transforms the shape model represented in terms of a voxel form into a triangular patch form, thirdly colors vertexes of triangular patches, and finally displays the shap...
متن کاملViewpoint invariant exemplar-based 3D human tracking
This paper proposes a clustered exemplar-based model for performing viewpoint invariant tracking of the 3D motion of a human subject from a single camera. Each exemplar is associated with multiple view visual information of a person and the corresponding 3D skeletal pose. The visual information takes the form of contours obtained from different viewpoints around the subject. The inclusion of mu...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE Access
سال: 2020
ISSN: 2169-3536
DOI: 10.1109/access.2020.3001230