Difference between revisions of "Private:progress-calagari"
From NMSL
(7 intermediate revisions by the same user not shown) | |||
Line 3: | Line 3: | ||
** CMPT-880: Large-scale Multimedia Systems and Cloud Computing | ** CMPT-880: Large-scale Multimedia Systems and Cloud Computing | ||
+ | ===July 16 - July 30 === | ||
+ | * Finalizing the course project | ||
+ | * Writing the report for the final project | ||
+ | |||
+ | ===July 2 - July 16 === | ||
+ | * Maneged to get a almost acceptable depth map using HD stereo videos and the @depth software. | ||
+ | * Choosing HOG for finding similar images and implementing it | ||
+ | * Using SIFT-flow for warping and implementing the previous work which uses a 3d data base for depth estimation. (but the parameters aren't tuned yet and the results are not as good as expected) | ||
+ | |||
+ | ===June 18 - July 2 === | ||
+ | * Course presentation | ||
+ | * Tried different softwares and codes for extracting the depth map using the two stereo images but non of them gave a good result. I even wrote a code for refining the depth map and tried to refine their results but still it wasn't good enough for the propose I needed :( | ||
+ | |||
+ | ===June 5 - June 18 === | ||
+ | * Working on making a data-set for my project: finding 3D football videos, partitioning it to frames and to the left and right image of each frame, using stereo matching programs to extract the depth out of these stereo images. | ||
+ | |||
+ | === May 21 - June 5 === | ||
+ | * Wrote a report about my idea: [https://cs-nsl-svn.cs.surrey.sfu.ca/cssvn/nsl-members/calagari/reports/Report-(DraftOfMyIdea).pdf report_link] | ||
+ | * Reading papers including : Tarek's report on depth signatures, the 2 papers Dr. Shervin sent me, and a few related papers. | ||
=== May 7 - May 21 === | === May 7 - May 21 === | ||
* Learning about SIFT flow as a warping tool that might be the missing piece for implementing our proposed method. | * Learning about SIFT flow as a warping tool that might be the missing piece for implementing our proposed method. | ||
− | * | + | * Studying the two improvements done on the data driven conversion approach. |
Latest revision as of 11:57, 14 August 2013
Summer 2012-2013 (RA)
- Courses:
- CMPT-880: Large-scale Multimedia Systems and Cloud Computing
July 16 - July 30
- Finalizing the course project
- Writing the report for the final project
July 2 - July 16
- Maneged to get a almost acceptable depth map using HD stereo videos and the @depth software.
- Choosing HOG for finding similar images and implementing it
- Using SIFT-flow for warping and implementing the previous work which uses a 3d data base for depth estimation. (but the parameters aren't tuned yet and the results are not as good as expected)
June 18 - July 2
- Course presentation
- Tried different softwares and codes for extracting the depth map using the two stereo images but non of them gave a good result. I even wrote a code for refining the depth map and tried to refine their results but still it wasn't good enough for the propose I needed :(
June 5 - June 18
- Working on making a data-set for my project: finding 3D football videos, partitioning it to frames and to the left and right image of each frame, using stereo matching programs to extract the depth out of these stereo images.
May 21 - June 5
- Wrote a report about my idea: report_link
- Reading papers including : Tarek's report on depth signatures, the 2 papers Dr. Shervin sent me, and a few related papers.
May 7 - May 21
- Learning about SIFT flow as a warping tool that might be the missing piece for implementing our proposed method.
- Studying the two improvements done on the data driven conversion approach.
- ===============================================================================
Spring 2012-2013 (RA)
- Courses:
- CMPT-820: Multimedia Systems
Apr 18 - May 2
- Figuring out a way that could be suitable for solving the search problem I had in my previous idea (2D-to-3D conversion of a football scene using 3D data sets)
- Reading papers about 2D-to-3D coversion
Apr 4 - Apr 18
- Figuring out a way that could be suitable for 2D-to-3D conversion of a football scene using 3D data sets.
- Reading papers about 2D-to-3D coversion
Mar 21 - Apr 4
- Implemented an idea for 2D-to-3D conversion (which is good for landscapes and images with perspective such as street views) as the course project.
Mar 7 - Mar 21
- Second course presentation
- Started working with edge detection and line detection algorithms as useful tools for my own research and also needed for detecting vanishing points and vanishing lines for my course project.
Feb 21 - Mar 7
- First course presentation
- Working on writing the journal paper and finishing its draft.
Feb 7 - Feb 21
- Course assignment (a simple implementation of a video codec such as H.264)
- Started writing the journal paper in any time left.
Feb 7
- My report: here.
- I presented in the Lab meeting
- ===============================================================================
Fall 2012-2013 (RA)
- Courses:
- CMPT-705: Algorithms
- CMPT-771: Internet Arch. and Protocols
Sep 14 - Sep 28
- Learned about latex and how to work with it and wrote a report about HEVC
- Read a few papers and web pages about HEVC
- Read Som's paper :"Cloud-based Scalable Multimedia Streaming" and had a discussion with him about it and also future works and challenges