Department of Electrical & Computer Engineering Signal and Image Laboratory (SaIL) The University of Arizona®

Current Research

Automatic Breast Segmentation of Multimodal MRI Images:  José A. Rosado-Toro
Breast density has been advocated as a risk factor for development of breast cancer. A new fat-water imaging technique called RAD-GRASE acquires images of the entire breast in a few minutes that can generate fat-fraction maps used to assess breast density. There is currently a lack of automated and high-throughput method for segmenting the breats from these MRI images. The aim of this research project is to develop a method to automatically segment the breast tissue in MR images and yield fat-water profiles of the region of interest.

Image and Video Inpainting:  Ding Ding, Sundaresh Ram
Inpainting is a process of restoring missing or damaged regions of an image. Inpainting algorithms can be roughly grouped into two different categories, PDE-based and exemplar-based algorithms, according to the size of region to be inpainted. PDE-based algorithms are capable of restoring small and narrow gaps of an image, but may blur the image when the gaps become broad. Exemplar-based algorithms provide good results for large object removal, but may fail to recover some structures (e.g., straight lines) in an image. Our goal is to develop improvements to these algorithms to obtain a plausible inpainting result both geometrically and texturally.

Multi-Polarization Fringe Projection Imaging for High Dynamic Range Objects:  Basel Salahieh
Traditional fringe-projection three-dimensional (3D) imaging techniques struggle to estimate the shape of high dynamic range (HDR) objects where detected fringes are of limited visibility. Moreover, saturated regions of specular reflections can completely block any fringe patterns, leading to lost depth information. We propose a multi-polarization fringe projection (MPFP) imaging technique that eliminates saturated points and enhances the fringe contrast by selecting the proper polarized channel measurements.

Object Detection and Tracking:  Rohit Chacko Philip, Xin Gao, Sundaresh Ram
Detection and tracking of an object in a video is a crucial and complex problem for applications such as video surveillance, traffic monitoring and tracking organs in medical images. There are many factors that make object detection and tracking a challenge task, such as large camera motion, low contrast between objects with backgrounds, illumination variation, clutter and occlusion. This research project is aimed at developing an automated system capable of detecting and tracking objects in low-resolution wide area imagery.

Rock Image Segmentation and Classification:  Ramya Malladi, Sundaresh Ram
Accurate rock size distribution is important for production blasting in order to control and minimize the overall production costs. They are typically measured by using sieves, which are time consuming and expensive, and do not provide information that can be used for online control or process improvement. In this research work we are aimed at developing a novel image segmentation algorithm capable of segmenting each rock particle separately in order to provide accurate rock size distribution.

Forecasting Solar Power Intermittency Using Ground-Based Sky Imaging:  Vijai Thottathil Jayadevan
Solar power utilization at the utility-scale is a Grand Challenge. A major problem is the intermittent output of solar power plants due to passing clouds and night time. Intermittency limits the adoption of solar power by utility companies and industry because they require reliable, predictable power generation. This research work aims at forecasting the cloud movements by tracking them using a ground-based sky camera, thereby predicting the intermittency.

Detection and Segmentation of 3-D Cell Nuclei:  Sundaresh Ram
Understanding the spatial organization of genes within the 3-D space of the nucleus is important for regulating gene expression. The biologists are interested in quantitative methods for studying nuclear organization. They currently lack automated and high-throughput methods for quantitative and qualitative global analysis of 3-D gene organization. The aim of this project is to build an automated system capable of detection, segmentation and classification of 3-D spots obtained from confocal microscopy.

Mid-Sagittal Tongue Detection and Tracking Using Ultrasound Images:  José A. Rosado-Toro
Ultrasound imaging is one of the most common imaging modalities used for study of the vocal tract. The advantage it has over X-ray and MRI is that, it is able to monitor the movement of the tongue due to its high frame-acquisition rate. Due to its high acquisition rate, manual detection becomes extremely tedious and time consuming. The aim of this research endeavor is to merge empirical properties of the tongue with machine learning algorithms in order to develop a robust and efficient detection/tracking system.

Tracking in Cardiac MR Images:  Srinivas L. Naik, Liangchh "James" Huang
The Study of Right Ventricular (RV) function has become very important recently for prognostic evaluation of several cardiac diseases, including chronic heart failure. Assesing the volume of right ventricle and its range of movement during a cardiac cycle are of atmost importance for the cardiologists in this study which is performed manually now. The main goal of this research work is to design and validate an automated system which will track the boundary of the RV region exactly.

Past Research

 1230 E. Speedway Blvd., P.O. Box 210104, Tucson, AZ 85721-0104
 ©2014 All Rights Reserved.  
 Contact webmaster                                  
Think ECE!