Apply your CUDA skills to image buffers by processing pixels on the GPU. You will map RGB data to threads, handle boundary conditions, and build a pipeline chaining grayscale conversion and Sobel edge detection. Finally, you will validate and export results, mastering real-world GPU image manipulation.
Overview
Syllabus
- Unit 1: RGB To Grayscale Conversion
- Seeing Brightness Correctly
- Guarding Every Pixel
- Finding Pixels in Memory
- Making Space on Gpu
- Checking the Final Result
- Pixels Into Shades
- Scaling Grayscale Images
- Unit 2: Implementing a Blur Filter
- Guarding Blur Edges
- Threads Into Pixels
- Building the Blur Window
- Blur Averaging Check
- Pixels in Memory
- Unit 3: Image Filter Pipeline
- Sharpening the Image Pipeline
- Sobel Edge Strength
- Keeping the Pipeline Alive
- Sobel Neighborhood Pass