A novel multi-atlas segmentation approach under the semi-supervised learning framework: Application to knee cartilage segmentation.

BACKGROUND AND OBJECTIVE: Multi-atlas based segmentation techniques, which rely on an atlas library comprised of training images labeled by an expert, have proven their effectiveness in multiple automatic segmentation applications. However, the usage of exhaustive patch libraries combined with the voxel-wise labeling incur a large computational cost in terms of memory requirements and execution times. METHODS: To confront this shortcoming, we propose a novel two-stage multi-atlas approach designed under the Semi-Supervised Learning (SSL) framework. The main properties of our method are as follows: First, instead of the voxel-wise labeling approach, the labeling of target voxels is accomplished here by exploiting the spectral content of globally sampled datasets from the target image, along with their spatially correspondent data collected from the atlases. Following SSL, voxels classification is boosted by incorporating unlabeled data from the target image, in addition to the labeled ones from atlas library. Our scheme integrates constructively fruitful concepts, including sparse reconstructions of voxels from linear neighborhoods, HOG feature descriptors of patches/regions, and label propagation via sparse graph constructions. Segmentation of the target image is carried out in two stages: stage-1 focuses on the sampling and labeling of global data, while stage-2 undertakes the above tasks for the out-of-sample data. Finally, we propose different graph-based methods for the labeling of global data, while these methods are extended to deal with the out-of-sample voxels. RESULTS: A thorough experimental investigation is conducted on 76 subjects provided by the publicly accessible Osteoarthritis Initiative (OAI) repository. Comparative results and statistical analysis demonstrate that the suggested methodology exhibits superior segmentation performance compared to the existing patch-based methods, across all evaluation metrics (DSC:88.89%, Precision: 89.86%, Recall: 88.12%), while at the same time it requires a considerably reduced computational load (>70% reduction on average execution time with respect to other patch-based). In addition, our approach is favorably compared against other non patch-based and deep learning methods in terms of performance accuracy (on the 3-class problem). A final experimentation on a 5-class setting of the problems demonstrates that our approach is capable of achieving performance comparable to existing state-of-the-art knee cartilage segmentation methods (DSC:88.22% and DSC:85.84% for femoral and tibial cartilage respectively).

Subject recognition based on ground reaction force measurements of gait signals.

An effective subject recognition approach is designed in this paper, using ground reaction force (GRF) measurements of human gait. The method is a three-stage procedure: 1) The original GRF data are translated through wavelet packet (WP) transform in the time-frequency domain. Using a fuzzy-set-based criterion, we determine an optimal WP decomposition, involving feature subspaces with distinguishing gait characteristics. 2) A feature extraction scheme is employed next for wavelet feature ranking, according to discrimination power. 3) The classification task is accomplished by means of a kernel-based support vector machine. The design parameters of the classifier are tuned through a genetic algorithm to improve recognition rates. The method is evaluated on a database comprising GRF records obtained from 40 subjects. To account for the natural variability of human gait, the experimental setup is designed, allowing different walking speeds and loading conditions. Simulation results demonstrate that high recognition rates can be achieved with moderate number of features and for different training/testing settings. Finally, the performance of our approach is favorably compared with the one obtained using other traditional classification algorithms.

Pipelined recurrent fuzzy neural networks for nonlinear adaptive speech prediction.

A class of pipelined recurrent fuzzy neural networks (PRFNNs) is proposed in this paper for nonlinear adaptive speech prediction. The PRFNNs are modular structures comprising a number of modules that are interconnected in a chained form. Each module is implemented by a small-scale recurrent fuzzy neural network (RFNN) with internal dynamics. Due to module nesting, the PRFNNs offer a number of desirable attributes, including decomposition of the modeling task, enhanced temporal processing capabilities, and multistage dynamic fuzzy inference. Tuning of the PRFNN adaptable parameters is accomplished by a series of gradient descent methods with different weighting of the modules and the decoupled extended Kalman filter (DEKF) algorithm, based on weight grouping. Extensive experimentation is carried out to evaluate the performance of the PRFNNs on the speech prediction platform. Comparative analysis shows that the PRFNNs outperform the single-RFNN models in terms of the prediction gains that are obtained and computational efficiency. Furthermore, PRFNNs provide considerably better performance compared to pipelined recurrent neural networks, for models with similar model complexity.

A dynamic fuzzy neural filter for separation of discontinuous adventitious sounds from vesicular sounds.

This paper presents a recurrent filter that performs real-time separation of discontinuous adventitious sounds from vesicular sounds. The filter uses two Dynamic Fuzzy Neural Networks, operating in parallel, to perform the task of separation of the lung sounds, obtained from patients with pulmonary pathology. Extensive experimental results, including fine/coarse crackles and squawks, are given, and a performance comparison with a series of other models is conducted, underlining the separation capabilities of the proposed filter and its improved performance with respect to its competing rivals.

A stable learning algorithm for block-diagonal recurrent neural networks: application to the analysis of lung sounds.

A novel learning algorithm, the Recurrent Neural Network Constrained Optimization Method (RENNCOM) is suggested in this paper, for training block-diagonal recurrent neural networks. The training task is formulated as a constrained optimization problem, whose objective is twofold: (1) minimization of an error measure, leading to successful approximation of the input/output mapping and (2) optimization of an additional functional, the payoff function, which aims at ensuring network stability throughout the learning process. Having assured the network and training stability conditions, the payoff function is switched to an alternative form with the scope to accelerate learning. Simulation results on a benchmark identification problem demonstrate that, compared to other learning schemes with stabilizing attributes, the RENNCOM algorithm has enhanced qualities, including, improved speed of convergence, accuracy and robustness. The proposed algorithm is also applied to the problem of the analysis of lung sounds. Particularly, a filter based on block-diagonal recurrent neural networks is developed, trained with the RENNCOM method. Extensive experimental results are given and performance comparisons with a series of other models are conducted, underlining the effectiveness of the proposed filter.