Project Background

Medical imaging has been around for a long time, but as the imaging instruments have improved, the focus has shifted away from generating the images towards analyzing them. Algorithmic techniques to inspect and extract information from medical images are constantly advancing. This project hopes to demonstrate a functional, straightforward, and usable implementation of such an algorithm. It is split into two components, each of which have been extensively researched: interpolation and segmentation.

Interpolation is used chiefly to generate data points inside the range of some set of input data points. Functionally, interpolators map a data set to a function which connects the dots and allows the user to use intermediate values between data points. In terms of imaging and computer vision, interpolators are used to scale an image to arbitrary size. Higher-order interpolators have a greater number of constraints leading to smoother functions - this may lead to a more accurate interpolated representation of the original image.

Segmentation then allows object recognition within the interpolated image, and subsequent analysis of the object in isolation from the rest of the image. Techniques such as thresholding and region growing can be employed to select an object from the image, filtering can be applied to remove noise, and then descriptions and features of the resulting object can be obtained.

More information about the specific interpolation and segmentation methods used can be found on their respective pages.

A video demonstrating a commonly used medical imaging technique: X-ray computed tomography scanning, also known as a CT scan.