Volume rendering is a procedure used to visualize two-dimensional projections of three-dimensional data sets. Paramount to this technique is the transfer function, which assigns optical properties to points inside the volume (e.g. color and transparency). By changing the transfer function, the user may choose what data to display or highlight. Selecting or changing the parameters of the transfer function is unintuitive and time-consuming for the average user. For such cases, using tools to generate transfer functions is helpful. In this work, we developed a model that generates transfer functions with the help of principal component analysis. A user interface, specialized in displaying and collecting transfer function parameters was developed. When a sufficient amount of data was obtained, it was analysed and restructured for further use. Then, the model was developed and installed into the VPT framework for volume visualization. With the update, users can now choose between suggested transfer functions, generated based on collected data.
|