Farzad Khalvati

EFIS—Evolving Fuzzy Image Segmentation

Despite the large number of techniques proposed in recent years, accurate segmentation of digital images remains a challenging task for automated computer algorithms. Approaches based on machine learning hold particular promise in this regard, because in many applications, e.g., medical image analysis, frequent user intervention can be assumed to correct the results, thereby generating valuable feedback for algorithmic learning. In order to learn segmentation of new (unseen) images, such user feedback (correction of current or past results) seems indispensable. In this paper, we propose the formation and evolution of fuzzy rules for user-oriented environments in which feedback is captured by design. The evolving fuzzy image segmentation (EFIS) can be used to adjust the parameters of existing segmentation methods, switch between their results, or fuse their results. Specifically, we propose a single-parametric EFIS (SEFIS), apply its rule evolution to breast ultrasound images, and evaluate the results using three segmentation methods, namely, global thresholding, region growing, and statistical region merging. The results show increased accuracy across all tests and for all methods. For instance, the accuracy of statistical region merging can be improved from 59% ± 30% to 71% ± 22%. We also propose a multiparametric EFIS (MEFIS) for switching between or fusing the results of multiple segmentation methods. Preliminary results indicate that MEFIS can further increase overall segmentation accuracy. Three thresholding methods with accuracies of 62% ± 11%, 64% ± 16%, and 61% ± 9% were combined to reach an overall accuracy of 66% ± 15%. Finally, we compare our SEFIS scheme with five other thresholding methods to evaluate its overall performance.

Medical Image Classification via SVM Using LBP Features from Saliency-Based Folded Data

Good results on image classification and retrieval using support vector machines (SVM) with local binary patterns (LBPs) as features have been extensively reported in the literature where an entire image is retrieved or classified. In contrast, in medical imaging, not all parts of the image may be equally significant or relevant to the image retrieval application at hand. For instance, in lung x-ray image, the lung region may contain a tumour, hence being highly significant whereas the surrounding area does not contain significant information from medical diagnosis perspective. In this paper, we propose to detect salient regions of images during training and fold the data to reduce the effect of irrelevant regions. As a result, smaller image areas will be used for LBP features calculation and consequently classification by SVM. We use IRMA 2009 dataset with 14,410 xray images to verify the performance of the proposed approach. The results demonstrate the benefits of saliency-based folding approach that delivers comparable classification accuracies with state-of-the-art but exhibits lower computational cost and storage requirements, factors highly important for big data analytics.

Autoencoding the retrieval relevance of medical images

Content-based image retrieval (CBIR) of medical images is a crucial task that can contribute to a more reliable diagnosis if applied to big data. Recent advances in feature extraction and classification have enormously improved CBIR results for digital images. However, considering the increasing accessibility of big data in medical imaging, we are still in need of reducing both memory requirements and computational expenses of image retrieval systems. This work proposes to exclude the features of image blocks that exhibit a low encoding error when learned by a n/p/n autoencoder (p <; n). We examine the histogram of autoendcoding errors of image blocks for each image class to facilitate the decision which image regions, or roughly what percentage of an image perhaps, shall be declared relevant for the retrieval task. This leads to reduction of feature dimensionality and speeds up the retrieval process. To validate the proposed scheme, we employ local binary patterns (LBP) and support vector machines (SVM) which are both well-established approaches in CBIR research community. As well, we use IRMA dataset with 14,410 x-ray images as test data. The results show that the dimensionality of annotated feature vectors can be reduced by up to 50% resulting in speedups greater than 27% at expense of less than 1% decrease in the accuracy of retrieval when validating the precision and recall of the top 20 hits.

A Multi-Parametric Diffusion Magnetic Resonance Imaging Texture Feature Model for Prostate Cancer Analysis

In this work, we present a new multi-parametric magnetic resonance imaging (MP-MRI) texture feature model for automatic detection of prostate cancer. In addition to commonly used imaging sequences in conventional MP-MRI, namely T2-weighted MRI (T2w) and diffusion-weighted imaging (DWI), our proposed MP-MRI texture feature model uses computed high-b DWI (CHB-DWI) and a new diffusion imaging sequence called correlated diffusion imaging (CDI). A set of texture features was calculated for both the conventional MP-MRI and new MP-MRI texture feature model. We evaluated the performance of the proposed MP-MRI texture feature model via leave-one-patient-out cross-validation using a Bayesian classifier trained on cancerous and healthy tissue samples obtained from real clinical MP-MRI datasets. The proposed MP-MRI texture feature model outperformed the conventional model (i.e., T2w+DWI) with regard to cancer detection accuracy.

Opposition-Based Window Memoization for Morphological Algorithms

In this paper we combine window memoization, a performance optimization technique for image processing, with opposition-based learning, a new learning scheme where the opposite of data under study is also considered in solving a problem. Window memoization combines memoization techniques from software and hardware with the repetitive nature of image data to reduce the number of calculations required for an image processing algorithm. We applied window memoization and opposition-based learning to a morphological edge detector and found that a large portion of the calculations performed on pixels neighborhoods can be skipped and instead, previously calculated results can be reused. The typical speedup for window memoization was 1.42. Combining window memoization with opposition-based learning yielded a typical increase of 5% in speedups

CT texture analysis: a potential tool for prediction of survival in patients with metastatic clear cell carcinoma treated with sunitinib

BackgroundTo assess CT texture based quantitative imaging biomarkers in the prediction of progression free survival (PFS) and overall survival (OS) in patients with clear cell renal cell carcinoma undergoing treatment with Sunitinib.MethodsIn this retrospective study, measurable lesions of 40 patients were selected based on RECIST criteria on standard contrast enhanced CT before and 2 months after treatment with Sunitinib. CT Texture analysis was performed using TexRAD research software (TexRAD Ltd, Cambridge, UK). Using a Cox regression model, correlation of texture parameters with measured time to progression and overall survival were assessed. Evaluation of combined International Metastatic Renal-Cell Carcinoma Database Consortium Model (IMDC) score with texture parameters was also performed.ResultsSize normalized standard deviation (nSD) alone at baseline and follow-up after treatment was a predictor of OS (Hazard ratio (HR) = 0.01 and 0.02; 95% confidence intervals (CI): 0.00 – 0.29 and 0.00 – 0.39; p = 0.01 and 0.01). Entropy following treatment and entropy change before and after treatment were both significant predictors of OS (HR = 2.68 and 87.77; 95% CI = 1.14 – 6.29 and 1.26 – 6115.69; p = 0.02 and p = 0.04). nSD was also a predictor of PFS at baseline and follow-up (HR = 0.01 and 0.01: 95% CI: 0.00 – 0.31 and 0.001 – 0.22; p = 0.01 and p = 0.003). When nSD at baseline or at follow-up was combined with IMDC, it improved the association with OS and PFS compared to IMDC alone.ConclusionSize normalized standard deviation from CT at baseline and follow-up scans is correlated with OS and PFS in clear cell renal cell carcinoma treated with Sunitinib.

Dual-stage correlated diffusion imaging

Diffusion imaging has become a powerful diagnostic imaging tool for the detection of prostate cancer. Recently, a new form of diffusion imaging called correlated diffusion imaging (CDI) was introduced, where the tissue being imaged is characterized by the joint correlation of diffusion signal attenuation across multiple gradient pulse strengths and timings. While CDI has been shown to provide a strong signal delineation between cancerous tissue and healthy tissue within the prostate gland, it does not capture anatomical information. As such, additional magnetic resonance imaging (MRI) modalities are required to be studied by the radiologists at the same time as CDI to provide anatomical context for accurate cancer localization within the prostate gland. In this study, we address this issue by introducing a new form of CDI called dual-stage correlated diffusion imaging (D-CDI), where an additional signal mixing stage is performed between the correlated diffusion signal from the first signal mixing stage and an auxiliary diffusion signal to incorporate anatomical context. Quantitative evaluation using sensitivity, specificity, and accuracy measures, and visual assessment by an expert radiologist for datasets of 13 patient cases with confirmed prostate cancer suggest that D-CDI not only provides strong delineation between healthy and cancerous tissues, it allows for accurate cancer localization in the prostate gland without the need for additional MRI modalities to be studied.

Automated Segmentation of Breast in 3-D MR Images Using a Robust Atlas

This paper presents a robust atlas-based segmentation (ABS) algorithm for segmentation of the breast boundary in 3-D MR images. The proposed algorithm combines the well-known methodologies of ABS namely probabilistic atlas and atlas selection approaches into a single framework where two configurations are realized. The algorithm uses phase congruency maps to create an atlas which is robust to intensity variations. This allows an atlas derived from images acquired with one MR imaging sequence to be used to segment images acquired with a different MR imaging sequence and eliminates the need for intensity-based registration. Images acquired using a Dixon sequence were used to create an atlas which was used to segment both Dixon images (intra-sequence) and T1-weighted images (inter-sequence). In both cases, highly accurate results were achieved with the median Dice similarity coefficient values of 94% ±4% and 87±6.5%, respectively.

Multiparametric MRI prostate cancer analysis via a hybrid morphological-textural model.

Multiparametric MRI has shown considerable promise as a diagnostic tool for prostate cancer grading. Diffusion-weighted MRI (DWI) has shown particularly strong potential for improving the delineation between cancerous and healthy tissue in the prostate gland. Current automated diagnostic methods using multiparametric MRI, however, tend to either use low-level features, which are difficult to interpret by radiologists and clinicians, or use highly subjective heuristic methods. We propose a novel strategy comprising a tumor candidate identification scheme and a hybrid textural-morphological feature model for delineating between cancerous and non-cancerous tumor candidates in the prostate gland via multiparametric MRI. Experimental results using clinical multiparametric MRI datasets show that the proposed strategy has strong potential as a diagnostic tool to aid radiologists and clinicians identify and detect prostate cancer more efficiently and effectively.

Combining Equivalence Verification and Completion Functions

This work presents a new method for verifying optimized register-transfer-level implementations of pipelined circuits. We combine the robust, yet limited, capabilities of combinational equivalence verification with the modular and composable verification strategy of completion functions. We have applied this technique to a 32-bit OpenRISC processor and a Sobel edge-detector circuit. Each case study required less than fifteen verification obligations and each obligation could be checked in less than one minute. We believe that our approach will be applicable to a large class of pipelines with in-order execution.