Journal Publications (Deep Learning)¶
LSTM Fully Convolutional Networks for Time Series Classification¶
Abstract
Fully convolutional neural networks (FCNs) have been shown to achieve the state-of-the-art performance on the task of classifying time series sequences. We propose the augmentation of fully convolutional networks with long short term memory recurrent neural network (LSTM RNN) sub-modules for time series classification. Our proposed models significantly enhance the performance of fully convolutional networks with a nominal increase in model size and require minimal preprocessing of the data set. The proposed long short term memory fully convolutional network (LSTM-FCN) achieves the state-of-the-art performance compared with others. We also explore the usage of attention mechanism to improve time series classification with the attention long short term memory fully convolutional network (ALSTM-FCN). The attention mechanism allows one to visualize the decision process of the LSTM cell. Furthermore, we propose refinement as a method to enhance the performance of trained models. An overall analysis of the performance of our model is provided and compared with other techniques.
Microaneurysm detection using fully convolutional neural networks¶
Abstract
Backround and Objectives: Diabetic retinopathy is a microvascular complication of diabetes that can lead to sight loss if treated not early enough. Microaneurysms are the earliest clinical signs of diabetic retinopa- thy. This paper presents an automatic method for detecting microaneurysms in fundus photographies.
Methods: A novel patch-based fully convolutional neural network with batch normalization layers and Dice loss function is proposed. Compared to other methods that require up to five processing stages, it requires only three. Furthermore, to the best of the authors’ knowledge, this is the first paper that shows how to successfully transfer knowledge between datasets in the microaneurysm detection domain.
Results: The proposed method was evaluated using three publicly available and widely used datasets: E- Ophtha, DIARETDB1, and ROC. It achieved better results than state-of-the-art methods using the FROC metric. The proposed algorithm accomplished highest sensitivities for low false positive rates, which is particularly important for screening purposes.
Microaneurysm detection using deep learning and interleaved freezing¶
Abstract
Diabetes affects one in eleven adults. Diabetic retinopathy is a microvascular complication of diabetes and the leading cause of blindness in the working-age population. Microaneurysms are the earliest clinical signs of diabetic retinopathy. This paper proposes an automatic method for detecting microaneurysms in fundus photographies. A novel patch-based fully convolutional neural network for detection of microaneurysms is proposed. Compared to other methods that require five processing stages, it requires only two. Furthermore, a novel network fine-tuning scheme called Interleaved Freezing is presented. This procedure significantly reduces the amount of time needed to re-train a network and produces competitive results. The proposed method was evaluated using publicly available and widely used datasets: E-Ophtha and ROC. It outperforms the state-of-the-art methods in terms of free-response receiver operatic characteristic (FROC) metric. Simplicity, performance, efficiency and robustness of the proposed method demonstrate its suitability for diabetic retinopathy screening applications.
Journal Publications (Algorithms Research / Machine Learning)¶
Parallel Quick Sort using Thread Pool Pattern¶
Abstract
Sorting algorithms, their implementations and their applications in modern computing necessitates improvements for sorting large data sets quickly and efficiently. This paper will analyze the performance of a multi-threaded quick sort implemented using the thread pool pattern. The analysis will be done by comparing the time required to sort various data sets and their memory constraints, against the native sorting implementations of the Dual Pivot Quicksort and Merge Sort using the Fork-Join framework in the Oracle Java 8 programming language. Analysis is done of the effect of different number of processor (cores) of the test machine, as well as the performance barrier due to the initial time taken to create “p” threads, p being the number of processors. This paper also analyzes the limitations of the inbuilt Java method Arrays.parallelSort() and how the proposed system overcomes this problem. Finally, it also discuss possible improvements to the proposed system to further improve its performance.
AdaSort: Adaptive Sorting using Machine Learning¶
Abstract
Sorting algorithms and their implementations in modern computing requires improvements in sorting large data sets effectively, both with respect to time and memory consumed. This paper is aimed at reviewing multiple adaptive sorting algorithms, on the basis of selection of an algorithm based on the characteristics of the data set. Machine Learning allows us to construct an adaptive algorithm based on the analysis of the experimental data. A review of algorithms designed using Systems of Algorithmic Algebra and Genetic Algorithms was performed. Both methods are designed to target different use cases. Systems of Algorithmic Algebra is a representation of pseudo code that can be converted to high level code using Integrated toolkit for Design and Synthesis of programs, while the Genetic Algorithm attempts to optimize its fitness function and generate the most successful algorithm.