| dc.identifier.uri | http://hdl.handle.net/11401/77816 | |
| dc.description.sponsorship | This work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree. | en_US |
| dc.format | Monograph | |
| dc.format.medium | Electronic Resource | en_US |
| dc.language.iso | en_US | |
| dc.publisher | The Graduate School, Stony Brook University: Stony Brook, NY. | |
| dc.type | Thesis | |
| dcterms.abstract | Unlike a computer-generated world as in a game or an animated world, we live in a real world with real objects which have measurable physical properties associated with them. With this in mind, we present here an acquisition system which can capture real scenes and their physical properties like geometry, color and material properties and use them to present the scene to the user in a way which is more useful than the raw scene itself. This project as the title suggests deals with the synthetic manipulation of real scenes captured from camera (specifically an RGBD camera). 3-D reconstruction of captured scene and relighting of synthetic geometry has been researched quite extensively but combining these two worlds is a relatively new chapter in the field of Computer Graphics. Since the information available from a regular color camera is hard to use for material acquisition to a high degree of accuracy, we will see how we can use an RGBD camera to perform this task. Our algorithm has direct applications in cultural heritage preservation, law enforcement, animated world content creation and other such similar fields. With our method we are able to scan real scenery and convert them into digital scenes/models with material properties and this has a big impact on how animated content is currently created. I significant portion of the artists work can be done by our method. With material editing in place it is even easier to modify the appearance of any asset (object) in the scanned scene according to the need. There has been relatively less work in the area of relighting of real scenes. Relighting of synthetic objects and gathering reflectance (parametric BRDF) models of objects from single photographic image has been tried before, but competitive work on reconstruction and relighting of real scenes is still in its nascent stage. This is particularly of interest since the Kinect camera is able to capture RGB as well as depth images of a scene. In this project we use the reconstructed geometry from RGBD camera to compute the partial BRDF or reflectance model of the scene. Since the BRDF of a material is high degree function and computing it fully turns out to be impractical for our purpose, we will stick to parametric models of scene objects. We will show how we can segment a scanned scene into real-world objects and extract their (partial) BRDFs. Once the partial BRDF of the scene is estimated, we perform relighting on the scene with different parameters and verify these results. Finally once the material acquisition is done on the scanned scene, we will show how we can use this information to perform material editing on the objects. We have included support for interactive material editing by the click of a mouse along with BRDF editing. Overall we present a system which addresses several challenges starting from scanning of real scenes and extracting their material properties to presenting them in a usable fashion to the user by enabling synthetic manipulability of the scene. | |
| dcterms.available | 2017-09-26T16:48:46Z | |
| dcterms.contributor | Qin, Hong | en_US |
| dcterms.contributor | Samaras, Dimitris. | en_US |
| dcterms.contributor | Mitchell, Joseph S.B. | en_US |
| dcterms.creator | Das, Ritwik | |
| dcterms.dateAccepted | 2017-09-26T16:48:46Z | |
| dcterms.dateSubmitted | 2017-09-26T16:48:46Z | |
| dcterms.description | Department of Computer Science. | en_US |
| dcterms.extent | 59 pg. | en_US |
| dcterms.format | Monograph | |
| dcterms.format | Application/PDF | en_US |
| dcterms.identifier | Das_grad.sunysb_0771M_12118.pdf | en_US |
| dcterms.identifier | http://hdl.handle.net/11401/77816 | |
| dcterms.issued | 2014-05-01 | |
| dcterms.language | en_US | |
| dcterms.provenance | Submitted by Jason Torre (fjason.torre@stonybrook.edu) on 2017-09-26T16:48:46Z
No. of bitstreams: 1
Das_grad.sunysb_0771M_12118.pdf: 28629697 bytes, checksum: 3e125f59b0a482c9a8e581ac16bc4f75 (MD5) | en |
| dcterms.provenance | Made available in DSpace on 2017-09-26T16:48:46Z (GMT). No. of bitstreams: 1
Das_grad.sunysb_0771M_12118.pdf: 28629697 bytes, checksum: 3e125f59b0a482c9a8e581ac16bc4f75 (MD5)
Previous issue date: 2014-05-01 | en |
| dcterms.publisher | The Graduate School, Stony Brook University: Stony Brook, NY. | |
| dcterms.subject | Kinect, Material Acquisition, Material Editing, Partial BRDF, Real Scene, Scene segmentation | |
| dcterms.subject | Computer science | |
| dcterms.title | REALight: A Material Acquisition and Editing System for Kinect Scenes | |
| dcterms.type | Thesis | |
