Application of StyleGAN Architecture for Generating Venous Leg Ulcer Images.

Leg ulcers caused by impaired venous blood return are the most typical chronic wound form and have a significant negative impact on the lives of people living with these wounds. Thus, it is important to provide early assessment and appropriate treatment of the wounds to promote their healing in the normal trajectory. Gathering quality wound data is an important component of good clinical care, enabling monitoring of healing progress. This data can also be useful to train machine learning algorithms with a view to predicting healing. Unfortunately, a high volume of good-quality data is needed to create datasets of suitable volume from people with wounds. In order to improve the process of gathering venous leg ulcer (VLU) data we propose the generative adversarial network based on StyleGAN architecture to synthesize new images from original samples. We utilized a dataset that was manually collected as part of a longitudinal observational study of VLUs and successfully synthesized new samples. These synthesized samples were validated by two clinicians. In future work, we plan to further process these new samples to train a fully automated neural network for ulcer segmentation.

On the inter-dataset generalization of machine learning approaches to Parkinson's disease detection from voice.

BACKGROUND AND OBJECTIVE: Parkinson's disease is the second-most-common neurodegenerative disorder that affects motor skills, cognitive processes, mood, and everyday tasks such as speaking and walking. The voices of people with Parkinson's disease may become weak, breathy, or hoarse and may sound emotionless, with slurred words and mumbling. Algorithms for computerized voice analysis have been proposed and have shown highly accurate results. However, these algorithms were developed on single, limited datasets, with participants possessing similar demographics. Such models are prone to overfitting and are unsuitable for generalization, which is essential in real-world applications. METHODS: We evaluated the computerized Parkinson's disease diagnosis performance of various machine learning models and showed that these models degraded rapidly when used on different datasets. We evaluated two mainstream state-of-the-art approaches, one based on deep convolutional neural networks and another based on voice feature extraction followed by a shallow classifier (i.e., extreme gradient boosting (XGBoost)). RESULTS: An investigation with four datasets (CzechPD, PC-GITA, ITA, and RMIT-PD) proved that even if the algorithms yielded excellent performance on a single dataset, the results obtained on new data or even a mix of datasets were very unsatisfactory. CONCLUSIONS: More work needs to be done to make computerized voice analysis methods for Parkinson's disease diagnosis suitable for real-world applications.

Bankruptcy prediction using ensemble of autoencoders optimized by genetic algorithm.

The prediction of imminent bankruptcy for a company is important to banks, government agencies, business owners, and different business stakeholders. Bankruptcy is influenced by many global and local aspects, so it can hardly be anticipated without deeper analysis and economic modeling knowledge. To make this problem even more challenging, the available bankruptcy datasets are usually imbalanced since even in times of financial crisis, bankrupt companies constitute only a fraction of all operating businesses. In this article, we propose a novel bankruptcy prediction approach based on a shallow autoencoder ensemble that is optimized by a genetic algorithm. The goal of the autoencoders is to learn the distribution of the majority class: going concern businesses. Then, the bankrupt companies are represented by higher autoencoder reconstruction errors. The choice of the optimal threshold value for the reconstruction error, which is used to differentiate between bankrupt and nonbankrupt companies, is crucial and determines the final classification decision. In our approach, the threshold for each autoencoder is determined by a genetic algorithm. We evaluate the proposed method on four different datasets containing small and medium-sized enterprises. The results show that the autoencoder ensemble is able to identify bankrupt companies with geometric mean scores ranging from 71% to 93.7%, (depending on the industry and evaluation year).

Feature Selection Based on a Sparse Neural-Network Layer With Normalizing Constraints.

Feature selection (FS) is an important step in machine learning since it has been shown to improve prediction accuracy while suppressing the curse of dimensionality of high-dimensional data. Neural networks have experienced tremendous success in solving many nonlinear learning problems. Here, we propose a new neural-network-based FS approach that introduces two constraints, the satisfaction of which leads to a sparse FS layer. We performed extensive experiments on synthetic and real-world data to evaluate the performance of our proposed FS method. In the experiments, we focus on high-dimensional, low-sample-size data since they represent the main challenge for FS. The results confirm that the proposed FS method based on a sparse neural-network layer with normalizing constraints (SNeL-FS) is able to select the important features and yields superior performance compared to other conventional FS methods.

Computerized analysis of speech and voice for Parkinson's disease: A systematic review.

BACKGROUND AND OBJECTIVE: Speech impairment is an early symptom of Parkinson's disease (PD). This study has summarized the literature related to speech and voice in detecting PD and assessing its severity. METHODS: A systematic review of the literature from 2010 to 2021 to investigate analysis methods and signal features. The keywords "Automatic analysis" in conjunction with "PD speech" or "PD voice" were used, and the PubMed and ScienceDirect databases were searched. A total of 838 papers were found on the first run, of which 189 were selected. One hundred and forty-seven were found to be suitable for the review. The different datasets, recording protocols, signal analysis methods and features that were reported are listed. Values of the features that separate PD patients from healthy controls were tabulated. Finally, the barriers that limit the wide use of computerized speech analysis are discussed. RESULTS: Speech and voice may be valuable markers for PD. However, large differences between the datasets make it difficult to compare different studies. In addition, speech analytic methods that are not informed by physiological understanding may alienate clinicians. CONCLUSIONS: The potential usefulness of speech and voice for the detection and assessment of PD is confirmed by evidence from the classification and correlation results.

Comparison of CNN-Learned vs. Handcrafted Features for Detection of Parkinson's Disease Dysgraphia in a Multilingual Dataset.

Parkinson's disease dysgraphia (PDYS), one of the earliest signs of Parkinson's disease (PD), has been researched as a promising biomarker of PD and as the target of a noninvasive and inexpensive approach to monitoring the progress of the disease. However, although several approaches to supportive PDYS diagnosis have been proposed (mainly based on handcrafted features (HF) extracted from online handwriting or the utilization of deep neural networks), it remains unclear which approach provides the highest discrimination power and how these approaches can be transferred between different datasets and languages. This study aims to compare classification performance based on two types of features: features automatically extracted by a pretrained convolutional neural network (CNN) and HF designed by human experts. Both approaches are evaluated on a multilingual dataset collected from 143 PD patients and 151 healthy controls in the Czech Republic, United States, Colombia, and Hungary. The subjects performed the spiral drawing task (SDT; a language-independent task) and the sentence writing task (SWT; a language-dependent task). Models based on logistic regression and gradient boosting were trained in several scenarios, specifically single language (SL), leave one language out (LOLO), and all languages combined (ALC). We found that the HF slightly outperformed the CNN-extracted features in all considered evaluation scenarios for the SWT. In detail, the following balanced accuracy (BACC) scores were achieved: SL-0.65 (HF), 0.58 (CNN); LOLO-0.65 (HF), 0.57 (CNN); and ALC-0.69 (HF), 0.66 (CNN). However, in the case of the SDT, features extracted by a CNN provided competitive results: SL-0.66 (HF), 0.62 (CNN); LOLO-0.56 (HF), 0.54 (CNN); and ALC-0.60 (HF), 0.60 (CNN). In summary, regarding the SWT, the HF outperformed the CNN-extracted features over 6% (mean BACC of 0.66 for HF, and 0.60 for CNN). In the case of the SDT, both feature sets provided almost identical classification performance (mean BACC of 0.60 for HF, and 0.58 for CNN).

Convolutional neural network ensemble for Parkinson's disease detection from voice recordings.

The computerized detection of Parkinson's disease (PD) will facilitate population screening and frequent monitoring and provide a more objective measure of symptoms, benefiting both patients and healthcare providers. Dysarthria is an early symptom of the disease and examining it for computerized diagnosis and monitoring has been proposed. Deep learning-based approaches have advantages for such applications because they do not require manual feature extraction, and while this approach has achieved excellent results in speech recognition, its utilization in the detection of pathological voices is limited. In this work, we present an ensemble of convolutional neural networks (CNNs) for the detection of PD from the voice recordings of 50 healthy people and 50 people with PD obtained from PC-GITA, a publicly available database. We propose a multiple-fine-tuning method to train the base CNN. This approach reduces the semantical gap between the source task that has been used for network pretraining and the target task by expanding the training process by including training on another dataset. Training and testing were performed for each vowel separately, and a 10-fold validation was performed to test the models. The performance was measured by using accuracy, sensitivity, specificity and area under the ROC curve (AUC). The results show that this approach was able to distinguish between the voices of people with PD and those of healthy people for all vowels. While there were small differences between the different vowels, the best performance was when/a/was considered; we achieved 99% accuracy, 86.2% sensitivity, 93.3% specificity and 89.6% AUC. This shows that the method has potential for use in clinical practice for the screening, diagnosis and monitoring of PD, with the advantage that vowel-based voice recordings can be performed online without requiring additional hardware.

Artificial Intelligence and Its Application to Minimal Hepatic Encephalopathy Diagnosis.

Hepatic encephalopathy (HE) is a brain dysfunction caused by liver insufficiency and/or portosystemic shunting. HE manifests as a spectrum of neurological or psychiatric abnormalities. Diagnosis of overt HE (OHE) is based on the typical clinical manifestation, but covert HE (CHE) has only very subtle clinical signs and minimal HE (MHE) is detected only by specialized time-consuming psychometric tests, for which there is still no universally accepted gold standard. Significant progress has been made in artificial intelligence and its application to medicine. In this review, we introduce how artificial intelligence has been used to diagnose minimal hepatic encephalopathy thus far, and we discuss its further potential in analyzing speech and handwriting data, which are probably the most accessible data for evaluating the cognitive state of the patient.

Selective oversampling approach for strongly imbalanced data.

Challenges posed by imbalanced data are encountered in many real-world applications. One of the possible approaches to improve the classifier performance on imbalanced data is oversampling. In this paper, we propose the new selective oversampling approach (SOA) that first isolates the most representative samples from minority classes by using an outlier detection technique and then utilizes these samples for synthetic oversampling. We show that the proposed approach improves the performance of two state-of-the-art oversampling methods, namely, the synthetic minority oversampling technique and adaptive synthetic sampling. The prediction performance is evaluated on four synthetic datasets and four real-world datasets, and the proposed SOA methods always achieved the same or better performance than other considered existing oversampling methods.

Dysgraphia detection through machine learning.

Dysgraphia, a disorder affecting the written expression of symbols and words, negatively impacts the academic results of pupils as well as their overall well-being. The use of automated procedures can make dysgraphia testing available to larger populations, thereby facilitating early intervention for those who need it. In this paper, we employed a machine learning approach to identify handwriting deteriorated by dysgraphia. To achieve this goal, we collected a new handwriting dataset consisting of several handwriting tasks and extracted a broad range of features to capture different aspects of handwriting. These were fed to a machine learning algorithm to predict whether handwriting is affected by dysgraphia. We compared several machine learning algorithms and discovered that the best results were achieved by the adaptive boosting (AdaBoost) algorithm. The results show that machine learning can be used to detect dysgraphia with almost 80% accuracy, even when dealing with a heterogeneous set of subjects differing in age, sex and handedness.

Small- and medium-enterprises bankruptcy dataset.

Bankruptcy prediction is a long-standing issue that receives significant attention of academic researchers and industry practitioners. Most of the papers on bankruptcy prediction focus on companies that are listed on the stock market, and there are only limited data for the rest of the companies. These companies, not indexed at any stock market, represent a significant part of the economy. The presented dataset consists of financial ratios of Slovak companies. There are 21 distinctive financial ratios which are available for three consecutive years prior to evaluation year in which companies may have filed for bankruptcy or not. The companies come from four different industries - agriculture, construction, manufacture, retail. We provide data for four consecutive years 2013-2016 for each industry. All companies are categorized as small-medium enterprises according to EU classification. Prediction performance results on this dataset are published in the research paper "Bankruptcy prediction for small- and medium-sized companies using severely imbalanced datasets" (Zoricák et al., 2019).

Evaluation of handwriting kinematics and pressure for differential diagnosis of Parkinson's disease.

OBJECTIVE: We present the PaHaW Parkinson's disease handwriting database, consisting of handwriting samples from Parkinson's disease (PD) patients and healthy controls. Our goal is to show that kinematic features and pressure features in handwriting can be used for the differential diagnosis of PD. METHODS AND MATERIAL: The database contains records from 37 PD patients and 38 healthy controls performing eight different handwriting tasks. The tasks include drawing an Archimedean spiral, repetitively writing orthographically simple syllables and words, and writing of a sentence. In addition to the conventional kinematic features related to the dynamics of handwriting, we investigated new pressure features based on the pressure exerted on the writing surface. To discriminate between PD patients and healthy subjects, three different classifiers were compared: K-nearest neighbors (K-NN), ensemble AdaBoost classifier, and support vector machines (SVM). RESULTS: For predicting PD based on kinematic and pressure features of handwriting, the best performing model was SVM with classification accuracy of Pacc=81.3% (sensitivity Psen=87.4% and specificity of Pspe=80.9%). When evaluated separately, pressure features proved to be relevant for PD diagnosis, yielding Pacc=82.5% compared to Pacc=75.4% using kinematic features. CONCLUSION: Experimental results showed that an analysis of kinematic and pressure features during handwriting can help assess subtle characteristics of handwriting and discriminate between PD patients and healthy controls.

Decision support framework for Parkinson's disease based on novel handwriting markers.

Parkinson's disease (PD) is a neurodegenerative disorder which impairs motor skills, speech, and other functions such as behavior, mood, and cognitive processes. One of the most typical clinical hallmarks of PD is handwriting deterioration, usually the first manifestation of PD. The aim of this study is twofold: (a) to find a subset of handwriting features suitable for identifying subjects with PD and (b) to build a predictive model to efficiently diagnose PD. We collected handwriting samples from 37 medicated PD patients and 38 age- and sex-matched controls. The handwriting samples were collected during seven tasks such as writing a syllable, word, or sentence. Every sample was used to extract the handwriting measures. In addition to conventional kinematic and spatio-temporal handwriting measures, we also computed novel handwriting measures based on entropy, signal energy, and empirical mode decomposition of the handwriting signals. The selected features were fed to the support vector machine classifier with radial Gaussian kernel for automated diagnosis. The accuracy of the classification of PD was as high as 88.13%, with the highest values of sensitivity and specificity equal to 89.47% and 91.89%, respectively. Handwriting may be a valuable marker as a diagnostic and screening tool.

Analysis of in-air movement in handwriting: A novel marker for Parkinson's disease.

BACKGROUND AND OBJECTIVE: Parkinson's disease (PD) is the second most common neurodegenerative disease affecting significant portion of elderly population. One of the most frequent hallmarks and usually also the first manifestation of PD is deterioration of handwriting characterized by micrographia and changes in kinematics of handwriting. There is no objective quantitative method of clinical diagnosis of PD. It is thought that PD can only be definitively diagnosed at postmortem, which further highlights the complexities of diagnosis. METHODS: We exploit the fact that movement during handwriting of a text consists not only from the on-surface movements of the hand, but also from the in-air trajectories performed when the hand moves in the air from one stroke to the next. We used a digitizing tablet to assess both in-air and on-surface kinematic variables during handwriting of a sentence in 37 PD patients on medication and 38 age- and gender-matched healthy controls. RESULTS: By applying feature selection algorithms and support vector machine learning methods to separate PD patients from healthy controls, we demonstrated that assessing the in-air/on-surface hand movements led to accurate classifications in 84% and 78% of subjects, respectively. Combining both modalities improved the accuracy by another 1% over the evaluation of in-air features alone and provided medically relevant diagnosis with 85.61% prediction accuracy. CONCLUSIONS: Assessment of in-air movements during handwriting has a major impact on disease classification accuracy. This study confirms that handwriting can be used as a marker for PD and can be with advance used in decision support systems for differential diagnosis of PD.