Two Dimensional and Gesture Based Medical Visualization Interface and Image Processing Methodologies to Aid and Diagnose of Lung Cancer

Abstract

This work aims to implement specific image processing methodologies to aid the appropriate diagnosis of lung cancer and provide a sterile environment for interacting with the medical data in operation theatre during treatment. Computer Tomography (CT) gives a good resolution axial slice image of the lung. The tumor’s raw CT data analysis is quite impossible as the tumor’s pixel characteristics would be approximately matching its neighboring pixels. Thus, a basic preprocessing algorithm is required to differentiate target from the background. It is processed with level set segmentation to extract tumor from the sequentially acquired multiple slices. The segmented output of each slice provides the geometric feature of the tumor. As the tumor size and shape in segmented portions are irregular, it seems appropriate to reconstruct a three-dimensional representation of 2D images for tumors’ qualitative information. The volume reconstruction using the ray casting method has been applied to render the volume of tumors from a segmented stack of 2D slices. Then touch-less, computer-aided, gesture-based control of the medical images was attained using Kinect sensor, which is the best tool for human-computer interaction to maintain a sterile environment. The paper elaborates on the gestures, which is employed to interact with the medical images such as selection, drag, and swipe gesture. The selection gesture is used to open and close individual medical images. The drag gesture was used to view the patient data along with the slice image. The swipe gesture used to view various medical data sets like 2D slices, tumor irregularity in segmented slices.