GPU-based relative fuzzy connectedness image segmentation

Parallel fuzzy minimals on GPU

Clustering is a classification method that organizes objects into groups based on their similarity. Data clustering can extract valuable information, such as human behavior, trends, and so on, from large datasets by using either hard or fuzzy approaches. However, this is a time-consuming problem due to the increasing volumes of data collected. In this context, sequential executions are not feasible and their parallelization is mandatory to complete the process in an acceptable time. Parallelization requires redesigning algorithms to take advantage of massively parallel platforms. In this paper we propose a novel parallel implementation of the fuzzy minimals algorithm on graphics processing unit as a high-performance low-cost solution for common clustering issues. The performance of this implementation is compared with an equivalent algorithm based on the message passing interface. Numerical simulations show that the proposed solution on graphics processing unit can achieve high performances with regards to the cost-accuracy ratio.

CUDA-based acceleration and BPN-assisted automation of bilateral filtering for brain MR image restoration

PURPOSE

Bilateral filters have been substantially exploited in numerous magnetic resonance (MR) image restoration applications for decades. Due to the deficiency of theoretical basis on the filter parameter setting, empirical manipulation with fixed values and noise variance-related adjustments has generally been employed. The outcome of these strategies is usually sensitive to the variation of the brain structures and not all the three parameter values are optimal. This article is in an attempt to investigate the optimal setting of the bilateral filter, from which an accelerated and automated restoration framework is developed.

METHODS

To reduce the computational burden of the bilateral filter, parallel computing with the graphics processing unit (GPU) architecture is first introduced. The NVIDIA Tesla K40c GPU with the compute unified device architecture (CUDA) functionality is specifically utilized to emphasize thread usages and memory resources. To correlate the filter parameters with image characteristics for automation, optimal image texture features are subsequently acquired based on the sequential forward floating selection (SFFS) scheme. Subsequently, the selected features are introduced into the back propagation network (BPN) model for filter parameter estimation. Finally, the k-fold cross validation method is adopted to evaluate the accuracy of the proposed filter parameter prediction framework.

RESULTS

A wide variety of T1-weighted brain MR images with various scenarios of noise levels and anatomic structures were utilized to train and validate this new parameter decision system with CUDA-based bilateral filtering. For a common brain MR image volume of 256 × 256 × 256 pixels, the speed-up gain reached 284. Six optimal texture features were acquired and associated with the BPN to establish a "high accuracy" parameter prediction system, which achieved a mean absolute percentage error (MAPE) of 5.6%. Automatic restoration results on 2460 brain MR images received an average relative error in terms of peak signal-to-noise ratio (PSNR) less than 0.1%. In comparison with many state-of-the-art filters, the proposed automation framework with CUDA-based bilateral filtering provided more favorable results both quantitatively and qualitatively.

CONCLUSIONS

Possessing unique characteristics and demonstrating exceptional performances, the proposed CUDA-based bilateral filter adequately removed random noise in multifarious brain MR images for further study in neurosciences and radiological sciences. It requires no prior knowledge of the noise variance and automatically restores MR images while preserving fine details. The strategy of exploiting the CUDA to accelerate the computation and incorporating texture features into the BPN to completely automate the bilateral filtering process is achievable and validated, from which the best performance is reached.

An improved parallel fuzzy connected image segmentation method based on CUDA

PURPOSE

Fuzzy connectedness method (FC) is an effective method for extracting fuzzy objects from medical images. However, when FC is applied to large medical image datasets, its running time will be greatly expensive. Therefore, a parallel CUDA version of FC (CUDA-kFOE) was proposed by Ying et al. to accelerate the original FC. Unfortunately, CUDA-kFOE does not consider the edges between GPU blocks, which causes miscalculation of edge points. In this paper, an improved algorithm is proposed by adding a correction step on the edge points. The improved algorithm can greatly enhance the calculation accuracy.

METHODS

In the improved method, an iterative manner is applied. In the first iteration, the affinity computation strategy is changed and a look up table is employed for memory reduction. In the second iteration, the error voxels because of asynchronism are updated again.

RESULTS

Three different CT sequences of hepatic vascular with different sizes were used in the experiments with three different seeds. NVIDIA Tesla C2075 is used to evaluate our improved method over these three data sets. Experimental results show that the improved algorithm can achieve a faster segmentation compared to the CPU version and higher accuracy than CUDA-kFOE.

CONCLUSIONS

The calculation results were consistent with the CPU version, which demonstrates that it corrects the edge point calculation error of the original CUDA-kFOE. The proposed method has a comparable time cost and has less errors compared to the original CUDA-kFOE as demonstrated in the experimental results. In the future, we will focus on automatic acquisition method and automatic processing.

Body-wide hierarchical fuzzy modeling, recognition, and delineation of anatomy in medical images

To make Quantitative Radiology (QR) a reality in radiological practice, computerized body-wide Automatic Anatomy Recognition (AAR) becomes essential. With the goal of building a general AAR system that is not tied to any specific organ system, body region, or image modality, this paper presents an AAR methodology for localizing and delineating all major organs in different body regions based on fuzzy modeling ideas and a tight integration of fuzzy models with an Iterative Relative Fuzzy Connectedness (IRFC) delineation algorithm. The methodology consists of five main steps: (a) gathering image data for both building models and testing the AAR algorithms from patient image sets existing in our health system; (b) formulating precise definitions of each body region and organ and delineating them following these definitions; (c) building hierarchical fuzzy anatomy models of organs for each body region; (d) recognizing and locating organs in given images by employing the hierarchical models; and (e) delineating the organs following the hierarchy. In Step (c), we explicitly encode object size and positional relationships into the hierarchy and subsequently exploit this information in object recognition in Step (d) and delineation in Step (e). Modality-independent and dependent aspects are carefully separated in model encoding. At the model building stage, a learning process is carried out for rehearsing an optimal threshold-based object recognition method. The recognition process in Step (d) starts from large, well-defined objects and proceeds down the hierarchy in a global to local manner. A fuzzy model-based version of the IRFC algorithm is created by naturally integrating the fuzzy model constraints into the delineation algorithm. The AAR system is tested on three body regions - thorax (on CT), abdomen (on CT and MRI), and neck (on MRI and CT) - involving a total of over 35 organs and 130 data sets (the total used for model building and testing). The training and testing data sets are divided into equal size in all cases except for the neck. Overall the AAR method achieves a mean accuracy of about 2 voxels in localizing non-sparse blob-like objects and most sparse tubular objects. The delineation accuracy in terms of mean false positive and negative volume fractions is 2% and 8%, respectively, for non-sparse objects, and 5% and 15%, respectively, for sparse objects. The two object groups achieve mean boundary distance relative to ground truth of 0.9 and 1.5 voxels, respectively. Some sparse objects - venous system (in the thorax on CT), inferior vena cava (in the abdomen on CT), and mandible and naso-pharynx (in neck on MRI, but not on CT) - pose challenges at all levels, leading to poor recognition and/or delineation results. The AAR method fares quite favorably when compared with methods from the recent literature for liver, kidneys, and spleen on CT images. We conclude that separation of modality-independent from dependent aspects, organization of objects in a hierarchy, encoding of object relationship information explicitly into the hierarchy, optimal threshold-based recognition learning, and fuzzy model-based IRFC are effective concepts which allowed us to demonstrate the feasibility of a general AAR system that works in different body regions on a variety of organs and on different modalities.