Learning 3D Models from a Single Still Image
Sign in to YouTube
Sign in to YouTube
Sign in to YouTube
Uploaded on Feb 2, 2008
Google Tech Talks
January, 29 2008
We present an algorithm to convert standard digital pictures into
This is a challenging problem, since an image is formed by a projection of the 3-d scene onto two dimensions, thus losing the depth information. We take a supervised learning approach to this problem, and use a Markov Random Field (MRF) to model the image depth cues as well as the relationships between different parts of the image. We show that even on unstructured scenes (of indoor and outdoor environments, including forests, trees, buildings,
etc.), our algorithm is frequently able to recover fairly accurate 3-d models.
We use our method to create visually pleasing 3-d flythroughs from the
image. We also present a few extensions of these ideas, such as additionally incorporating triangulation (stereo) cues, and using multiple images to produce large scale 3-d models. We also apply our methods to two robotics applications: (a) high speed offroad obstacle avoidance on an autonomously driven remote-controlled car, and (b) having a robot unload items from a dishwasher.
To convert your own image of an outdoor scene, landscape, etc. to a 3-d model, please visit: http://make3d.stanford.edu
Joint work with Min Sun and Andrew Y. Ng.
Speaker: Ashutosh Saxena
Ashutosh is a PhD candidate with Prof. Andrew Y. Ng in the Computer
Science department in Stanford University. He received his B. Tech.
from Indian Institute of Technology (IIT Kanpur) in 2004.
His research focuses on machine learning approaches to problems in
computer vision and in robotic manipulation. Using data-driven machine
learning techniques, he developed algorithms for creating 3-d models from
a single image, and algorithms for robotic manipulation tasks such as
opening doors, and grasping previously unseen objects.
Standard YouTube License
- 59:23 The Next Generation of Neural Networksby GoogleTechTalks 276,098 views
- 58:53 Urban Reconstruction and Modeling for Building Virtual Worldsby GoogleTechTalks 14,264 views
- 57:25 Visual Perception with Deep Learningby GoogleTechTalks 31,444 views
- 52:51 Simple interactive 3D modeling for allby GoogleTechTalks 39,039 views
- 1:02:35 Visual 3D modeling of real-world objects and scenes from...by GoogleTechTalks 7,229 views
- 55:27 How Cybercriminals Steal Moneyby GoogleTechTalks 134,832 views
- 1:13:40 Quantum Computing Day 2: Image Recognition with an Adiabatic Quantum Computerby GoogleTechTalks 36,267 views
- 1:05:18 Recent Developments in Deep Learningby GoogleTechTalks 47,358 views
- 23 videos Play all Google Tech Talksby Daniro88
- 58:08 No Time to Thinkby GoogleTechTalks 131,286 views
- 1:13:07 Lecture 4 | Machine Learning (Stanford)by StanfordUniversity 74,056 views
- 1:01:20 The Lively Kernelby GoogleTechTalks 48,496 views
- 56:33 PhotoSketch: A Photo-Centric 3D Modeling Toolby GoogleTechTalks 7,394 views
- 55:44 Optimization for Machine Learningby GoogleTechTalks 10,749 views
- 1:22:14 Lecture: Probabilistic Methods in Computer Systems Modelingby USCViterbi 16,853 views
- 48:40 Speech recognition and retrieval using unsupervised sub-word language modelsby GoogleTechTalks 9,835 views
- 9:50 PhotoSketch Plugin for Google SketchUp: Introductionby PhotoSketchPlugin 29,463 views
- 9:13 PhotoAnim 3DX Tutorial #2 of 7 - Real photo to 3D Modelby PhotoAnim1 2,112 views
- 54:57 Algorithms for Data Management and Migrationby GoogleTechTalks 13,585 views
- 59:14 Hidden Topic Markov Modelsby GoogleTechTalks 19,140 views
- 1:15:31 Lecture 5 | Machine Learning (Stanford)by StanfordUniversity 54,062 views
- Loading more suggestions...