About Me

Ju Lynn Ong

Ph. D. Student
The Australian National University, Canberra & National ICT Australia

Research Area

My research is in the field of shape modelling and image processing, with applications in polyp detection from CT colonography datasets. Colorectal carcinoma or colon cancer is the second leading cause of cancer deaths in Australia, with approximately 4700 deaths/year. Studies so far have shown that this type of cancer often arises from pre-existent adenomatous polyps which grow from small adenomas (<5mm), into large adenomas (>10mm), then into non invasive carcinoma and finally to invasive carcinoma. The time taken for small adenomas to develop into cancer is 10-15 years on average, and studies have shown that early detection and removal of colonic polyps have resulted in a decline in mortality of about 76-90%. For this reason, checking for these precursors to colonic cancer is highly recommended. Computed Tomographic (CT) Colonography is a non-invasive technique that produces cross-sectional CT data scans used to reconstruct volumetric colonic data of the patient for further processing and analysis. Several CAD systems have been developed to aid the detection of these adenomatous polyps in order to be removed in the early stage. However, polyp detection is a challenging process because they come in various shapes and sizes and residual materials and folds in the colonic wall may also resemble polyps.

Overview of Research

Movies

Slice by slice scroll through 2D slices. Polyps are marked with a green circle. 2D Slices

Flythrough of 3D reconstructed colon. Polyps are highlighted in red. 3D Flythrough

Conference and Workshops

J.L. Ong and A-K. Seghouane, "Geodesic-Ring Based Curvature Maps for Polyp Detection in CT Colonography," in 2010 IEEE International Conference on Image Processing ICIP, Sept 2010.

J.L. Ong and A-K. Seghouane, "Efficient Feature Selection for Polyp Detection," in 2010 IEEE International Conference on Image Processing ICIP, Sept 2010.

A-K. Seghouane and J.L Ong, "Multiresolution Colonic Polyp Detection in CT Colonography Using Spherical Wavelets," in 2009 IEEE Workshop on Statistical Signal Processing, Wales, Aug 2009.

J.L. Ong, A.-K. Seghouane, "False Positive Reduction in CT Colonography using Spectral Compression and Curvature Tensor Smoothing of Surface Geometry," in Proc. IEEE International Symposium on Biomedical Imaging (ISBI), Boston, June 2009.

J.L. Ong, A.-K. Seghouane and K. Osborn, "Mean Shape Models for Polyp Detection in CT Colonography," in Digital Image Computing Techniques and Applications (DICTA), pp. 287-293, Canberra, Dec 2008.

J.L. Ong, A.-K. Seghouane, "A Method for Classification of Candidate Lesions in CT Colonography," in Proc. of the MICCAI Workshop on Computational and Visualization Challenges in the New Era of Virtual Colonoscopy, pp.135-140, New York, Sept 2008.

J.L. Ong, A.-K. Seghouane and K. Osborn, "Polyp Detection in CT Colonography based on Shape Characteristics and Kullback-Leibler Divergence," in Proc. IEEE International Symposium on Biomedical Imaging (ISBI), Paris, May 2008.

Journal Publications

J.L. Ong and A.-K Seghouane, "From Point to Local Neighbourhood: Polyp Detection in CT Colonography using Geodesic Ring Neighbourhoods," accepted in IEEE Transactions on Image Processing, to appear.

J.L. Ong and A.-K Seghouane, "Feature Selection using Mutual Information in CT Colonography," accepted in Pattern Recognition Letters, to appear.