Full Wave Inversion (FWI) imaging scheme has many applications in engineering, geoscience and medical sciences. In this paper, a surrogate deep learning FWI approach is presented to quantify properties of materials using stress waves. Such inverse problems, in general, are ill-posed and nonconvex, especially in cases where the solutions exhibit shocks, heterogeneity, discontinuities, or large gradients. The proposed approach is proven efficient to obtain global minima responses in these cases. This approach is trained based on random sampled set of material properties and sampled trials around local minima, therefore, it requires a forward simulation can handle high heterogeneity, discontinuities and large gradients. High resolution Kurganov-Tadmor (KT) central finite volume method is used as forward wave propagation operator. Using the proposed framework, material properties of 2D media are quantified for several different situations. The results demonstrate the feasibility of the proposed method for estimating mechanical properties of materials with high accuracy using deep learning approaches.

Heart disease is the leading cause of death worldwide. Amongst patients with cardiovascular diseases, myocardial infarction is the main cause of death. In order to provide adequate healthcare support to patients who may experience this clinical event, it is essential to gather supportive evidence in a timely manner to help secure a correct diagnosis. In this article, we study the feasibility of using deep learning to identify suggestive electrocardiographic (ECG) changes that may correctly classify heart conditions using the Physikalisch-Technische Bundesanstalt (PTB) database. As part of this study, we systematically quantify the contribution of each ECG lead to correctly tell apart a healthy from an unhealthy heart. For such a study we fine-tune the ConvNetQuake neural network model, which was originally designed to identify earthquakes. Our findings indicate that out of 15 ECG leads, data from the v6 and vz leads are critical to correctly identify myocardial infarction. Based on these findings, we modify ConvNetQuake to simultaneously take in raw ECG data from leads v6 and vz, achieving $99.43\%$ classification accuracy, which represents cardiologist-level performance level for myocardial infarction detection after feeding only 10 seconds of raw ECG data to our neural network model. This approach differs from others in the community in that the ECG data fed into the neural network model does not require any kind of manual feature extraction or pre-processing.

Purpose: To develop a fast, accurate, and robust convolutional neural network (CNN) based method for segmentation of thalamic nuclei. Methods: A cascaded multi-planar scheme with a modified residual U-Net architecture was used to segment thalamic nuclei on clinical datasets acquired using the white-matter-nulled Magnetization Prepared Rapid Gradient Echo (MPRAGE) sequence. A single network was optimized for healthy controls and disease types (multiple sclerosis, essential tremor) and magnetic field strengths (3T and 7T). Another network was developed to use conventional MPRAGE data. Clinical utility was assessed by comparing a cohort of MS patients to healthy subjects. Results: Segmentation of each thalamus into 12 nuclei was achieved in under 4 minutes. For 7T WMn-MPRAGE, the proposed method outperformed current state-of-the-art with statistically significant improvements in Dice ranging from 1.2% to 5.3% for MS and from 2.6% to 38.8% for ET patients. Comparable accuracy (Dice/VSI) was achieved between 7T and 3T data, attesting to the robustness of the method. For conventional MPRAGE, Dice of > 0.7 was achieved for larger nuclei and > 0.6 for the smaller nuclei. Atrophy of five thalamic nuclei and the whole thalamus was observed for MS patients compared to healthy control subjects, after controlling for intracranial volume and age (p<0.004). Conclusion: The proposed segmentation method is fast, accurate, and generalizes across disease types and field strengths and shows great potential for improving our understanding of thalamic nuclei involvement in neurological diseases and healthy aging. KEYWORDS Deep learning, convolutional neural network, transfer learning, thalamic nuclei segmentation

Within sEMG-based gesture recognition, a chasm exists in the literature between offline accuracy and real-time usability of a classifier. This gap mainly stems from the four main dynamic factors in sEMG-based gesture recognition: gesture intensity, limb position, electrode shift and transient changes in the signal. These factors are hard to include within an offline dataset as each of them exponentially augment the number of segments to be recorded. On the other hand, online datasets are biased towards the sEMG-based algorithms providing feedback to the participants, limiting the usability of such datasets as benchmarks. This paper proposes a virtual reality (VR) environment and a real-time experimental protocol from which the four main dynamic factors can more easily be studied. During the online experiment, the gesture recognition feedback is provided through the leap motion camera, enabling the proposed dataset to be re-used to compare future sEMG-based algorithms. 20 able-bodied persons took part in this study, completing three to four sessions over a period spanning between 14 and 21 days. Finally, TADANN, a new transfer learning-based algorithm, is proposed for long term gesture classification and significantly (p<0.05) outperforms fine-tuning a network.

For several skin conditions such as vitiligo, accurate segmentation of lesions from skin images is the primary measure of disease progression and severity. Existing methods for vitiligo lesion segmentation require manual intervention. Unfortunately, manual segmentation is time and labor-intensive, as well as irreproducible between physicians. We introduce a convolutional neural network (CNN) that quickly and robustly performs vitiligo skin lesion segmentation. Our CNN has a U-Net architecture with a modified contracting path. We use the CNN to generate an initial segmentation of the lesion, then refine it by running the watershed algorithm on high-confidence pixels. We train the network on 247 images with a variety of lesion sizes, complexity, and anatomical sites. The network with our modifications noticeably outperforms the state-of-the-art U-Net, with a Jaccard Index (JI) score of 73.6% (compared to 36.7%). Moreover, our method requires only a few seconds for segmentation, in contrast with the previously proposed semi-autonomous watershed approach, which requires 2-29 minutes per image.

Speech separation refers to extracting each individual speech source in a given mixed signal. Recent advancements in speech separation and ongoing research in this area, have made these approaches as promising techniques for pre-processing of naturalistic audio streams. After incorporating deep learning techniques into speech separation, performance on these systems is improving faster. The initial solutions introduced for deep learning based speech separation analyzed the speech signals into time-frequency domain with STFT; and then encoded mixed signals were fed into a deep neural network based separator. Most recently, new methods are introduced to separate waveform of the mixed signal directly without analyzing them using STFT. Here, we introduce a unified framework to include both spectrogram and waveform separations into a single structure, while being only different in the kernel function used to encode and decode the data; where, both can achieve competitive performance. This new framework provides flexibility; in addition, depending on the characteristics of the data, or limitations of the memory and latency can set the hyper-parameters to flow in a pipeline of the framework which fits the task properly. We extend single-channel speech separation into multi-channel framework with end-to-end training of the network while optimizing the speech separation criterion (i.e., Si-SNR) directly. We emphasize on how tied kernel functions for calculating spatial features, encoder, and decoder in multi-channel framework can be effective. We simulate spatialized reverberate data for both WSJ0 and LibriSpeech corpora here, and while these two sets of data are different in the matter of size and duration, the effect of capturing shorter and longer dependencies of previous/+future samples are studied in detail. We report SDR, Si-SNR and PESQ to evaluate the performance of developed solutions.

Driver vigilance estimation is an important task for transportation safety. Wearable and portable brain-computer interface devices provide a powerful means for real-time monitoring of the vigilance level of drivers, thus help with avoiding distracted or impaired driving. In this paper, we propose a novel multimodal architecture for in-vehicle vigilance estimation from Electroencephalogram and Electrooculogram. To tackle this problem and other issues associated with multimodal biological signal analysis, we propose an architecture composed of a capsule attention mechanism following a deep Long Short-Term Memory (LSTM) network. Our model learns both temporal and hierarchical/spatial dependencies in the data through the LSTM and capsule feature representation layers. To better explore the discriminative ability of the learned representations, we study the effect of the proposed capsule attention mechanism including the number of dynamic routing iterations as well as other parameters.

In vivo diffusion tensor imaging (DTI) is a promising technique to investigate noninvasively the fiber structures of the in vivo human heart. However, signal loss due to motions remains a persistent problem in in vivo cardiac DTI. We propose a novel motion-compensation method for investigating in vivo myocardium structures in DTI with free-breathing acquisitions. The method is based on an invertible Wavelet Scattering achieved by means of Convolutional Neural Network (WSCNN). It consists of first extracting translation-invariant wavelet scattering features from DW images acquired at different trigger delays and then mapping the fused scattering features into motion-compensated spatial DW images by performing an inverse wavelet scattering transform achieved using CNN. The results on both simulated and acquired in vivo cardiac DW images showed that the proposed WSCNN method effectively compensates for motion-induced signal loss and produces in vivo cardiac DW images with better quality and more coherent fiber structures with respect to existing methods, which makes it an interesting method for measuring correctly the diffusion properties of the in vivo human heart in DTI under free breathing.

This paper investigates the intriguing question of whether we can create learning algorithms that automatically generate training data, learning environments, and curricula in order to help AI agents rapidly learn. We show that such algorithms are possible via Generative Teaching Networks (GTNs), a general approach that is, in theory, applicable to supervised, unsupervised, and reinforcement learning, although our experiments only focus on the supervised case. GTNs are deep neural networks that generate data and/or training environments that a learner (e.g. a freshly initialized neural network) trains on for a few SGD steps before being tested on a target task. We then differentiate through the entire learning process via meta-gradients to update the GTN parameters to improve performance on the target task. GTNs have the beneficial property that they can theoretically generate any type of data or training environment, making their potential impact large. This paper introduces GTNs, discusses their potential, and showcases that they can substantially accelerate learning. We also demonstrate a practical and exciting application of GTNs: accelerating the evaluation of candidate architectures for neural architecture search (NAS), which is rate-limited by such evaluations, enabling massive speed-ups in NAS. GTN-NAS improves the NAS state of the art, finding higher performing architectures when controlling for the search proposal mechanism. GTN-NAS also is competitive with the overall state of the art approaches, which achieve top performance while using orders of magnitude less computation than typical NAS methods. Speculating forward, GTNs may represent a first step toward the ambitious goal of algorithms that generate their own training data and, in doing so, open a variety of interesting new research questions and directions.

In this paper we present ensembles of classifiers for automated animal audio classification, exploiting different data augmentation techniques for training Convolutional Neural Networks (CNNs). The specific animal audio classification problems are i) birds and ii) cat sounds, whose datasets are freely available. We train five different CNNs on the original datasets and on their versions augmented by four augmentation protocols, working on the raw audio signals or their representations as spectrograms. We compared our best approaches with the state of the art, showing that we obtain the best recognition rate on the same datasets, without ad hoc parameter optimization. Our study shows that different CNNs can be trained for the purpose of animal audio classification and that their fusion works better than the stand-alone classifiers. To the best of our knowledge this is the largest study on data augmentation for CNNs in animal audio classification audio datasets using the same set of classifiers and parameters. Our MATLAB code is available at https://github.com/LorisNanni .