This page lists papers submitted for IJFIPM that have been reviewed and
accepted but not yet published. Please note: article titles, authors,
abstracts and keywords may change upon publication.
Our TOC e-mail alerting service will notify you immediately when new issues
of IJFIPM are published on-line. Click here to register for our TOC E-Mail Alerting. We also offer the convenience of RSS
feeds which provide a means to view new content timely posted to your web
site or desktop. Click here to start to use our free RSS news feeds.
IEEE 7th BIBE Special Issue Papers in Press
| 1. Efficient Super Granular SVM Feature Elimination (Super GSVM-FE)
Model for Protein Sequence Motif Information Extraction |
| Authors: |
Bernard Chen, Stephen Pellicer, Phang C. Tai, Robert Harrison and
Yi Pan |
| Abstract: |
Protein sequence motifs are gathering more and more attention in the sequence analysis area. The conserved regions have the potential to determine the conformation, function and activities of the proteins. In our previous work, we tried to obtain protein sequence motifs which are universally conserved and across protein family boundaries. Therefore, unlike most popular motif discovering algorithms, our input dataset is extremely large. In order to deal with huge amount of input dataset, we provided two granular computing models (FIK and FGK model) to efficiently generate protein motifs information. In this article, we develop a new method which combines the concept of granular computing and the power of Ranking-SVM to further extract protein sequence motif information. There are two reasons for eliminating redundant datasets: First, the information we try to generate is about sequence motif, but the original input data is derived from whole protein sequences by sliding window technique. Second, during fuzzy c-means clustering, it has the ability to assign one segment to more than one information granule. However, not all data segments have a direct relation to the granule to which they are assigned. The quality of motif information increases dramatically in all three evaluation measures by applying this new feature elimination model. Since the training step of Ranking-SVM is very time consuming, we also provide a feasible way to reduce the training time dramatically without sacrificing the quality. A new presentation of protein sequence motif which transcend protein family boundaries is also provided. |
| Keywords: |
FIK Model, FGK Model, Ranking SVM, Feature Elimination, Protein Sequence Motif |
| |
| 2. A Kernelized Fuzzy-Support Vector Machine CAD System for the Diagnosis of Lung Cancer from Tissue Images |
| Authors: |
Walker H. Land, Jr, Daniel W. McKee, Tatyana Zhukov, Dansheng Song, and Wei Qian |
| Abstract: |
This research describes a non-interactive process that applies several forms of computational intelligence to the task of classifying biopsy lung tissue samples based on visual data in the form of raw digital photographs of those samples. The three types of lung cancer evaluated (squamous cell carcinoma, adenocarcinoma, and bronchioalveolar carcinoma) together account for 65-70% of lung cancer diagnoses. The accuracy of the process on the test data supports the hypothesis that an accurate predictive model can be generated from the training images, and that the performance achieved in this study is an accurate baseline for the process’s potential performance against much larger quantities of data. Specifically, we demonstrate that the performance of our feature vector is as good as or better than Thiran and Macq’s in every case. With the exception of bronchioalveolar carcinomas, each individual cancer classification task experienced at least a modest improvement, with two groupings showing nearly 20% improvements in classification accuracy due to our feature vector. |
| Keywords: |
CAD, lung cancer, segmentation, feature selection, classification, microscopy images, kernel methods, SVM |
| |
| 3. Using fractal dimension as discriminator of infected HeLa cells from spectrophotometric images |
| Authors: |
Radu Dobrescu, Loretta Ichim and Stefan S. Nicolau |
| Abstract: |
The paper presents an original experimental optical method to characterize using fractal dimension, cell nuclei size distribution for virus infected and non-infected cells. There is described the solution to design an optical system which allow backscattering Mie diffusion spectra determination for biological sample on transparent holder. Fractal dimension for interest area of Mie scattering spectra was computed using a software package, which can store and process recorded data. The results indicate obviously higher values of fractal dimension for HSV virus infected biological samples compared to non-infected biological samples. This allows us clearly discriminate between virus infected and non-infected biological samples. |
| Keywords: |
Light scattering spectroscopy; spectra, signal analysis; Mie scattering; fractal dimension; cell nuclei. |
| |
| 4. An Efficient Compression Method for Multiplanar Reformulated Biomedical Images |
| Authors: |
Qiang Cheng and Mehdi Zargham |
| Abstract: |
Multiplanar reformatting (MPR) of 3D biomedical images is an important technique in visualizing, editing, and interacting volumetric data. To obtain near real-time interactions with the data in many applications such as telemedicine and teleconsultation, the MPR slices are produced and transmitted dynamically. The MPR compression is an important technique to improve the transmission and display efficiency. We develop a dedicated MPR compression scheme by exploiting the characteristics ofMPR slices, especially for thin MPR. Robust regression techniques are applied to predict the current slice from the previous ones. In the presence of scales, rotations, and translations, we make use of the known knowledge of the operations, or the transform domain representations of the image in the case of unknown parameters. The effectiveness of the scheme is confirmed by experimental results. |
| Keywords: |
MPR compression, biomedical images |
| |
| 5. Visualizing Menisci-Femur Contact Using Deformable Knee Models |
| Author: |
Ying Zhu |
| Abstract: |
Computational simulation of knee joint can help expand our understanding of the knee biomechanics and improve orthopedic practice. So far few computational knee models include a deformable menisci model. In addition, patient specific modeling and simulation are still not widely adopted, although they are shown to be more effective than generic modeling. In this paper, we present two new methods for simulating and visualizing menisci-femur contact area during gait cycles, using patient specific knee models and motion data. Specifically we propose a new template based 3D model reconstruction method that generates patient specific knee models -- including menisci ?from medical images. We also propose a new deformable menisci model for real-time applications. Our deformable model combines physics-based deformation and spatial deformation, making it easier for users to balance the physical realism and performance needs. These new methods will benefit knee biomechanical analysis as well as interactive orthopedic surgery planning and training. |
| Keywords: |
knee, model, simulation, visualization, software, image |
| |
| 6. Abnormalities of the Corpus Callosum in Autism Subtype |
| Authors: |
Qing He, Ye Duan, Judith Miles, and Nicole Takahashi |
| Abstract: |
Brain imaging studies of the corpus callosum (CC) in autism have yield inconsistent results. In this paper, we explore the three-dimensional profile of CC abnormalities in autism. The CC is segmented from mid-sagittal MRI and four adjacent slices on both sides, using our newly developed semiautomatic method. A subsequent contour stitching is performed to create the 3D surface of the CC, and the point correspondence problem can be simplified by our segmentation scheme. After alignment, differences between autistic patients and the control group are computed using Hotelling T2 two-sample metric, which results in a significance map. Permutation test is performed for multiple comparison and both raw and corrected p-values are shown in the results. Additional visualization of the group difference is provided via mean difference map. The statistical results reveal significant difference between patients and controls in the body of the CC. |
| Keywords: |
corpus callosum, autism, abnormalities, shape analysis, permutation test |
| |
| 7. New Statistical Learning Theory Paradigms Adapted to Breast Cancer Diagnosis / Classification Using Image and Non-Image Clinical Data |
| Authors: |
Walker H. Land, Jr, John Heine, Tom Raway, Alda Mizaku Nataliya Kovalchuk, Jack Y. Yang and Mary Qu Yang |
| Abstract: |
This paper describes three separate breast cancer research studies using both magnetic resonance mammography and screen film mammography coupled with clinical and other feature variables that addresses the following research question : how do we address the false positive biopsy artifact in diagnostic mammography, while still maintaining high sensitivities? Two new machine intelligence paradigms are used to study this problem: a Evolutionary Programming / Evolutionary Strategies Stochastic SVM hybrid (EP/ ES stochastic hybrid) and two kernalized Partial Least Squares paradigms (auto K-PLS and k-PLS). The research studies performed were: (1.) performance comparison of the EP /ES SVM stochastic hybrid with the standard iterative method of training using an identical data set and statistical cross validation methods, (2.) performance tradeoff and diagnostic accuracy of the auto-K-PLS, K-PLS and the EP / ES SVM hybrids using the MRM and non-image data sets, and (3.)developing EP /ES SVM hybrid and K-PLS classification accuracies using the BIRADS and clinical feature data set. These studies showed that:(1.) the new EP /ES hybrid produced comparable results to the more standard methods of iterative SVM training, but quicker, and (2.) the new auto K-PLS and K-PLS systems will train and operate in essentially real time for data sets of reasonable size. These faster training and operating times are advantages in the clinical environment. Specific breast cancer classification and diagnostic research results are discussed in the paper. |
| Keywords: |
Kernel-Partial Least Squares, Evolutionary Programming / Evolutionary Strategies Derived Support Vector Machines, Machine Intelligence, computer aided diagnosis / classification |
|
