<p>In this thesis, we reexamine the classical problems of image/video spatial resolution up conversion and video deinterlacing with an aim to develop real-time, adaptive solutions. The research of this thesis is important because most video applications require real time throughput. We study the use of GPU (Graphics Processing Unit) technology for high throughput video interpolation and deinterlacing. The main technical challenge is how to fully utilize the processing power and parallel architecture of GPU to maximize the throughput of up conversion and deinterlacing without compromising the visual quality of the resulting videos. To achieve the goal we develop a GPU-friendly two-pass directional image/video resolution up conversion algorithm and present a GPU implementation of the method, using the NVIDIA CUDA (Compute Unified Device Architecture) technology. We also devise a GPU-motivated motion-adaptive deinterlacing algorithm and develop a CUDA-based implementation of the algorithm. To strike a balance between performance and complexity, we discuss the techniques of adapting the computations in motion detection and adaptive directional interpolation to the GPU architecture for maximum video throughput possible. Experimental results demonstrate that using a mid-range GPU card, our CUDA-based implementations offer real-time solutions for image/video spatial resolution upconversion and video deinterlacing.</p> / Master of Applied Science (MASc)
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/9394 |
| Date | 05 1900 |
| Creators | Cao, Jie |
| Contributors | Wu, Xiaolin, Electrical and Computer Engineering |
| Source Sets | McMaster University |
| Detected Language | English |
| Type | thesis |
Page generated in 0.0017 seconds