Image Processing for Medical Applications

Diagnostic System Steps

• The classification step involves recognizing an object and making a yes/no decision, and then sorting objects into one of several groups or classes.
• Classification techniques can be divided into two broad areas: statistical or structural (or syntactic) techniques, with a third area that borrows from both, called cognitive methods.

Classification Techniques

• Statistical techniques deal with objects or patterns that have an underlying statistical basis for their generation and are described by quantitative features such as length, area, and texture.
• Structural (or syntactic) techniques deal with objects best described by qualitative features describing structural or syntactic relationships inherent in the object.
• Statistical classification methods are more popular than structural methods, and hybrid, cognitive methods have gained popularity over the last decade.

Connected Components Labeling

• Segmentation provides a simplified, binary image that separates objects of interest (foreground) from the background.
• The foreground pixels are set to “1,” and the background pixels set to “0.”
• Connected components labeling scans the segmented, binary image and groups its pixels into components based on pixel connectivity.

Connected Components Labeling Algorithm

• The algorithm assigns provisional labels to pixels based on their neighbors, and then replaces each label with the label assigned to its equivalence class in a second scan.
• A total of 7 labels are generated in accordance with the conditions highlighted above, and the label equivalence relationships are generated.

Feature Recognition and Classification

• Pre-processing step involves enhancing the quality of the acquired image, which depends on the resolution, sensitivity, bandwidth, and signal-to-noise ratio of the imaging system.
• Segmentation step involves isolating different objects from each other and from the background, and labeling the different objects.
• Feature extraction step reduces the data by measuring certain properties or features of the labeled objects, and these features are then passed to a classifier that evaluates the evidence presented.

Mean Shift Algorithm

• Mean Shift is a clustering algorithm that does not require the number of clusters as input.
• The algorithm only takes one input, the bandwidth of the window.
• Disadvantages include high computational cost, and the selection of the bandwidth itself can be non-trivial.