3DV Tutorial
3D Object Geometry from Single Image
in association with 3DV 2016
October 28th, Stanford University

Course Description

In the past decade, we have seen a revolution in 3D object recognition techniques from geometric approaches with handcrafted features to learning based approaches with data-driven representations. With the fast development of deep learning approaches and increasing availability of datasets, recovering 3D geometry of objects from a single image become a possible task. In this course, we will talk about the state-of-the-art techniques for 3D object category detection, pose estimation, keypoint localization, shape reconstruction as well as 3D human pose estimation. We will have hands-on opportunities using datasets and open source code.


Part I. We are going to discuss how to recover the 3D pose and structure of an object or a human body with semantic keypoints and a collection of 3D models. We will discuss the deformable shape model and the sparse representation for 3D structure recovery and related optimization techniques. We will also introduce keypoint detection in images with the state-of-the-art convolutional neural networks (CNNs) and how to address the uncertainties in the outputs of CNNs when inferring the 3D object geometry from them.

Part II. We are going to present the state-of-the-art techniques in 3D object category recognition, with emphasis on recognizing 3D properties of objects from single input images such as 3D pose, 3D location and 3D shape of objects. Especially, we are going to discuss methods and algorithms that employ 3D CAD models to help 3D object category recognition. We are going to introduce a large scale dataset for 3D object recognition named ObjectNet3D, which consists of 100 categories, 90,127 images, 201,888 objects in these images and 44,147 3D shapes. We will present how the dataset is constructed, and how the objects in the images are aligned with 3D shapes. In addition, we will discuss a 3D object recognition challenge using the ObjectNet3D dataset.



Organizers and Speakers


Xiaowei Zhou, postdoctoral researcher, University of Pennsylvania

Yu Xiang, postdoctoral researcher, University of Washington

Kostas Daniilidis, professor, University of Pennsylvania



Schedule
Date: October 28th
14:00 - 15:20 Part I
15:20 - 15:40 Coffee break
15:40 - 17:00 Part IIĀ 

Venue

Frost Amphitheater

Frances C. Arrillaga Alumni Center
326 Galvez Street
Stanford, CA 94305-6105

Materials
Slides for Part I: PDF
Slides for part II: PDF