Summary and next steps

These include:

• Types of image noise and noise reduction
• The use of filtering in noise reduction: Gaussian, mean and median filtering
• Contrast enhancement methods such as histogram equalisation
• Counting and labelling via image segmentation:, with the three main types being classification, clustering and spatial grouping
• Pixel classification via Otsu thresholding
• Clustering via K-means clustering
• Spatial grouping via split and merge using quadtrees, and region growing methods
• Feature-based methods such as edge and corner detection, and the Sobel and Canny operators
• Model-based image segmentation using active contours or snakes

So far, most of the image data we’ve discussed relate to single 2D images. Next week however, we look beyond 2D images and consider data types consisting of multiple images, such as video and 3D volumetric data.