Anaglyph Videos Student:Jihaad Pienaar Supervisor: Mr Mehrdad Ghaziasgar Co-Supervisor:Mr James Connan Mentors: Mr Roland Foster & Mr Kenzo Abrahams
Overview RECAP DESIGN DECISIONS AND SYSTEM CHANGES IMPLEMENTATION TOOLS USED PROJECT PLAN DEMO
Recap A system which allows users to convert 2D videos into an Anaglyph 3D equivalent.
Recap We were able to convert a stereoscopic 2D image into a 3D equivalent using an algorithm by Eric Dubois.
Design Decisions And System Changes
Disseminate video into frames. Use a depth map to produce left and right frame. Feed left and right frame into Dubois algorithm to produce 3D frame. Put frames back together to create 3D video
Implementation
Get the frame
Implementation Each frame we want to produce a depth map
Implementation Foreground extraction 4 th order moments
Implementation Fill holes
Implementation Determine background depth by creating a similar image but for blurred regions.
Implementation Generate stereoscopic pair Left Right
Tools Used
Ubuntu C/C++ OpenCV 2.4
Project Plan
Completed Gather Requirements and Analysis Whilst learning Image processing techniques in Open CV as well as researching 3D in general Completed Design and development of the prototype Prototype should be able to convert video into 3D equivalent. Completed Coding the final application Term 4 Testing and Evaluation
Demo
Questions
References Bradski, G., & Kaehler, A., “Learning OpenCV. United States of America: O’Reilly Media, Inc.”, pp Changick, K., Jungwoo, P., Jaeho, L. & Jenq-Neng, H., Fast Extraction of Objects of Interest from Images with Low Depth of Field. ETRI, 29(3), pp Jaeseung, K., Manbae, K. & Changick, K., D-To-3D Stereoscopic Conversion: Depth-Map Estimation in a 2D Single- View Image. Applications of Digital Image Processing XXX, pp. 1-9.