A web management service applied to a comprehensive characterization of Visible Human Dataset colour images

Abstract

Visible Human Dataset (VHD) is a remarkable piece of raw digital anatomical knowledge still to be fully exploited. Colours of VHD anatomic images are the natural targets of different algorithmic approaches devoted to understanding the content of the complex digital medical images, but they have never been analysed exhaustively. A full colorimetric characterization of all 9000 VHD colour images may help to take advantage of implicit available information in raw data. This study describes a novel colorimetric characterization and a Visual Knowledge Discovery tool, using methods from database field, data visualization, and image analysis. The applied heterogeneous methods allowed us to develop a histogram meta database and make it available remotely. It consists of a histogram-based colorimetric characterization of the all VHD 24-bit colour images. A user-friendly, interactive, and intuitive 3D framework providing 3D services was built and made freely available. It allows real-time analysis of colour component characteristics of a user-defined set of VHD images providing 3D interactive navigation of the histogram meta database. New knowledge can be discovered using our tool and the histogram meta database provided. This work allowed us to propose novel methods for colour image characterization and obtained results using developed service on VHD colour images let us to partially understand the not fully satisfactorily results achieved so far analysing these images.