Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


GPU Vision

Survey of computer vision algorithms on the GPU.

Umar Arshad

on 28 April 2010

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of GPU Vision

Computer Vision Pattern Recognition Feature Extraction Scene Reconstruction Optical Character Recognition Computer Vision on GPU Lecturer: Umar Arshad Instructor: Dr. Sushil Prasad Class: Parallel Algorithms GPU Graphics Why Massively Parallel Architecture
240 Scalar Programmable Cores
141.7 GB/s Memory Bandwidth ATI HD 5870
320 Vector Programming Cores
153.6 GB/s Memory Bandwidth Image String Image 3D Model Images = 512 512 262,144 Pixels * 30 FPS = 7,864,320 Pixel/s
IDEO? 7.5 MB/s Search Face Detection Navigation Programming Frameworks CUDA OpenCL OpenGL Close To Metal BrookGPU Conclusion How? Computer Vision is computationally intensive
GPU are an excellent platform to perform computer vision calculations
GPU based implementation are seeing speedups of 110x
Questions? SURF SIFT Viola and Jones Algorithm GLOH
Full transcript