CANCER 3D Modeling from 2D DICOM Images
I am a student in a university with a health background and some basic information in Python. I am working in a cancer visualization and extraction project and looking for an experienced software programmer using python and have experience in medical imaging as DICOM images (digital imaging and communication in medicine).This project continues for one month and through it I want the programmer to make 3D reconstruction from 2D DICOM images. The program also makes extraction and segmentation of the tumor tissue from the background and segmentation of other tissues. Then it renders the 3D model of the extracted tumor.
We need to use different kinds of algorithms as marching cube algorithm including interpolation and approximation for 3D reconstruction. Developer may use segmentation algorithms as K-means, region growing, threshold and the likes.
I already started the work and I have online resources for DICOM images, I performed some scripts related to this work and downloaded some open-source of DICOM images software processing images as 3D slicer and imageJ. All of these I will provide the programmer to start his/her work.
A program is needed to be written in Python. It will do the following:
1. Read sequence of DICOM .DCM images (2D) and render them in sequence and the sequence can be controlled with the mouse back and forth. (Already completed)
2. Segmenting and analyzing the 2D DICOM images and separate areas that are likely to be cancer (We will give the programmer some algorithms recommendations as well as the density that are likely to be cancer), separate cancer areas in new sequence of DICOM images, example:
a. Original DICOM images (1.dcm, 2.dcm, 3.dcm, 4.dcm, 5.dcm, 6.dcm, 7.dcm, 8.dcm, 9.dcm, 10.dcm, 11.dcm, 12.dcm).
b. The DICOM that contains cancer would be: (3.dcm, 4.dcm, 5.dcm, 11.dcm, 12.dcm).
c. The following requirement in number (3), it will build the 3D model from the cancer DICOM images in (b) above.
3. Then build a 3D construction from the DICOM sequence. The 3D model should be controlled with degrees of depth (From skin to the deepest structure). (Completed partly)
Figure 1: (2D DICOM image) Shows cancer in blue circle, the cancer can be analyzed according to several algorithms, some of them consider the brightness, some the circular shape, and some the density. The developer can use any algorithm that sees fits, to determine the cancer area in DICOM images.