Projects

Medical Imaging and Radiology Informatics (MIRI)

Medical Imaging Information System (MIIS)

Parallel Computing for Medical Imaging (PCMI)

Modeling of Biotransport in Biological Systems (MBBS)

Optical Imaging Tomography and Applications (OITA)

Image Stereology and Clinic Applications (ISCA)

Stereotactic Atlas for the Anatomic Topology (SAAT)

Coupled Diffusions for Image Enhencement (DDIE)

Grid-based Medical Imaging System (GMIS)

Cyberinfrastructure for Medical Imaging Informatics (CMII)

Medical Image Informatics Education (MIIE)

Nanomedicine and nanobiomedical imaging (NMNI)

Grant Opportunities

Description

Coupled Duffuions for Image Enhancement (DMIE)

Yongqiang Zhao, Ph.D., Shanghai Jiaotogn University, China and Department of Radiology, University of Iowa, USA, Minglu Li, Ph.D., Shanghai Jiaotogn University, China, Jun Ni, Ph.D. , Department of Radiology, University of Iowa, USA

Based on the recently-developed crease enhancement diffusion (CED) filter, and combined the sharpening qualities of shock filters with excellent edge preservation and excellent smoothing for flat regions of total variation (TV) filter, an adaptive couple diffusion (ACD) filter is to enhance the local coherence of vascular structures for medical images. This filter is tested on several experiments with synthetic and clinical images. The experimental results show the filter provides better performance for enhancement with further sharpening and de-noising features.

Based on the CED filter, we introduce a model composed of two nonlinear diffusion processes which offers an advantages in enhancement of image quality. The model is accomplished by replacing the Gaussian smoothed image with an adaptive coefficient that locally controls the sharpness and smoothness of the image. The advantages of the model are that non-vessel features and image contrast are enhanced while noisy background is suppressed and artifacts in the image are not created. The algorithm can be applied to filtering various medical images for significant improvement of image contrast.

.

.

.

.

.

.