Select one of the above methods for your experiments. You should refer to the appropriate section of the Tina documentation for this technique before continuing.
Using methods you have already learned on this course, estimate the amount of image noise in the input images. Remind yourself of how to generate random noise images with a similar quantity of white noise.
Take the original image data, and by adding suitable amounts of random noise (to all input images), determine the change in output data caused by additional random noise similar to the measured image noise.
Using the scatterplot methods construct histograms of measured value against the change in that value seen due to noise (Bland-Altmann plots). You will need to remember to scale data appropriately, this should be done so that the scaling on each axis is known.
Summarise and explain your findings, illustrating with suitable images and figures in a report of up to 5 pages.
Try to answer the following questions;
What are typical expected levels of noise in the output data as a percentage of the computed quantity?
Does random uniform noise produce uniform noise on output quantities?
How might the behaviour of data limit the use of these outputs?