Department Publications

Diagnosing pre-existing heart diseases early in life is important as it helps prevent complications such as pulmonary hypertension, heart rhythm problems, blood clots, heart failure and sudden cardiac arrest. To identify such diseases, phonocardiogram (PCG) and electrocardiogram (ECG) waveforms convey important information. Therefore, effectively using these two modalities of data has the potential to improve the disease screening process. We evaluate this hypothesis on a subset of the PhysioNet Challenge 2016 Dataset which contains simultaneously acquired PCG and ECG recordings. Our novel dual-convolutional neural network based approach uses transfer learning to tackle the problem of having limited amounts of simultaneous PCG and ECG data that is publicly available, while having the potential to adapt to larger datasets. In addition, we introduce two main evaluation frameworks named record-wise …

We present a MUSIC-based Direction of Arrival (DOA) estimation strategy using small antenna arrays, via employing deep learning for reconstructing the signals of a virtual large antenna array. Not only does the proposed strategy deliver significantly better performance than simply plugging the incoming signals into MUSIC, but surprisingly, the performance is also better than directly using an actual large antenna array with MUSIC for high angle ranges and low test SNR values. We further analyze the best choice for the training SNR as a function of the test SNR, and observe dramatic changes in the behavior of this function for different angle ranges.

Drones can be very useful in search and rescue missions primarily due to the aerial imaging capability. Drones can assist ground teams searching for a missing person, and law enforcement officers in crowd control. But there are instances where the observer on the ground control station misses the subtle information on the video feed coming from the drone. In fact, it is a very difficult job for the observer to be so vigilant in viewing a lot of images looking for a sign of a human in those images. This paper presents a drone-based human identification system to make search and rescue missions more effective. The drone, once detects signs of human presence in real-time aerial videos, autonomously navigates towards the suspicious location to get a better view to verify human presence. The drone has been made fully autonomous with operator override as an option. It processes images and sends selected frames for …

Convolutional Neural Networks (CNNs) have achieved state-of-the-art performance in many computer vision tasks over the years. However, this comes at the cost of heavy computation and memory intensive network designs, suggesting potential improvements in efficiency. Convolutional layers of CNNs partly account for such an inefficiency, as they are known to learn redundant features. In this work, we exploit this redundancy, observing it as the correlation between convolutional filters of a layer, and propose an alternative approach to reproduce it efficiently. The proposed `LinearConv' layer learns a set of orthogonal filters, and a set of coefficients that linearly combines them to introduce a controlled redundancy. We introduce a correlation-based regularization loss to achieve such flexibility over redundancy, and control the number of parameters in turn. This is designed as a plug-and-play layer to conveniently …

Learning a particular task from a dataset, samples in which originate from diverse contexts, is challenging, and usually addressed by deepening or widening standard neural networks. As opposed to conventional network widening, multipath architectures restrict the quadratic increment of complexity to a linear scale. However, existing multi-column/path networks or model ensembling methods do not consider any feature-dependent allocation of parallel resources, and therefore, tend to learn redundant features. Given a layer in a multi-path network, if we restrict each path to learn a context-specific set of features and introduce a mechanism to intelligently allocate incoming feature maps to such paths, each path can specialize in a certain context, reducing the redundancy and improving the quality of extracted features. This eventually leads to better-optimized usage of parallel resources. To do this, we propose …

Real-time and online action localization in a video is a critical yet highly challenging problem. Accurate action localization requires the utilization of both temporal and spatial information. Recent attempts achieve this by using computationally intensive 3D CNN architectures or highly redundant two-stream architectures with optical flow, making them both unsuitable for real-time, online applications. To accomplish activity localization under highly challenging real-time constraints, we propose utilizing fast and efficient key-point based bounding box prediction to spatially localize actions. We then introduce a tube-linking algorithm that maintains the continuity of action tubes temporally in the presence of occlusions. Further, we eliminate the need for a two-stream architecture by combining temporal and spatial information into a cascaded input to a single network, allowing the network to learn from both types of information. Temporal information is efficiently extracted using a structural similarity index map as opposed to computationally intensive optical flow. Despite the simplicity of our approach, our lightweight end-to-end architecture achieves state-of-the-art frame-mAP of 74.7% on the challenging UCF101-24 dataset, demonstrating a performance gain of 6.4% over the previous best online methods. We also achieve state-of-the-art video-mAP results compared to both online and offline methods. Moreover, our model achieves a frame rate of 41.8 FPS, which is a 10.7% improvement over contemporary real-time methods.

Self-localization based on passive RFID-based has many potential applications. One of the main challenges it faces is the suppression of the reflected signals from unwanted objects (i.e., clutter). Typically, the clutter echoes are much stronger than the backscattered signals of the passive tag landmarks used in such scenarios. Therefore, successful tag detection can be very challenging. We consider two types of tags, namely low-Q and high-Q tags. The high-Q tag features a sparse frequency response, whereas the low-Q tag presents a broad frequency response. Further, the clutter usually showcases a short-lived response. In this work, we propose an iterative algorithm based on a low-rank plus sparse recovery approach (RPCA) to mitigate clutter and retrieve the landmark response. In addition to that, we compare the proposed approach with the well-known time-gating technique. It turns out that RPCA outperforms significantly time-gating for low-Q tags, achieving clutter suppression and tag identification when clutter encroaches on the time-gating window span, whereas it also increases the backscattered power at resonance by approximately 12 dB at 80 cm for high-Q tags. Altogether, RPCA seems a promising approach to improve the identification of passive indoor self-localization tag landmarks.

Defect detection plays an important role in product quality assurance in many industrial applications. Air-coupled ultrasound defect detection setups have the advantage that they are non-destructive and do not contaminate the investigated materials. However, this approach comes with some challenges due to the weak signaling response of the defects. Thus, sophisticated signal separation methods are required. To address this challenge, low-rank-plus-sparse recovery (LRPSR) is proposed and compared with state-of-the-art machine learning (ML) methods. The results obtained show that the LRPSR method achieves comparable results in terms of detection rate to those achieved by ML. Yet, for a small training data set, the LRPSR approach outperforms the ML algorithms. The small training data set is important for detecting defects of different production lines without having a time-consuming training process. In …

Deep neural networks (DNNs) have been successfully employed in a multitude of applications with remarkable performance. As such performance is achieved at a significant computational cost, several embedded applications demand fast and efficient hardware accelerators for DNNs. Previously proposed application specific integrated circuit (ASIC) architectures strive to utilize arrays of hundreds of processing elements (PEs) and reduce power-hungry DRAM accesses using multiple dataflows requiring complex PE architectures. These consume significant area and reduce the maximum clock frequency. This paper introduces the Kraken architecture, which optimally processes the convolutional layers, fully-connected layers, and matrix products of any DNN through a hardware-friendly uniform dataflow. This enables maximal data reuse of weights, inputs, and outputs, with a bare-bones PE design and on-the-fly dynamic reconfiguration. Kraken, implemented in 65-nm CMOS technology at 400 MHz, packs 672 PEs in 7.3 mm2, with a peak performance of 537.6 Gops. Kraken processes the convolutional layers of AlexNet, VGG-16, and ResNet-50 at 336.6, 17.5, and 64.2 frames/s, respectively, hence outperforming the state-of-the-art ASIC architectures in terms of overall performance efficiency, DRAM accesses, arithmetic intensity, and throughput, with 5.8x more Gops/mm2 and 1.6x more Gops/W.

Image generation from a single image using generative adversarial networks is quite interesting due to the realism of generated images. However, recent approaches need improvement for such realistic and diverse image generation, when the global context of the image is important such as in face, animal, and architectural image generation. This is mainly due to the use of fewer convolutional layers for mainly capturing the patch statistics and, thereby, not being able to capture global statistics very well. We solve this problem by using attention blocks at selected scales and feeding a random Gaussian blurred image to the discriminator for training. Our results are visually better than the state-of-the-art particularly in generating images that require global context. The diversity of our image generation, measured using the average standard deviation of pixels, is also better.

Large-batch training of deep neural networks (DNN) has recently been widely studied, since traversing the optimization landscape is faster with large batches and the emergence of parallel computing has made large-batch training feasible. However, its generalization is still worse compared to small-batch training, which is commonly attributed to the low gradient variance with large batches. We show that sign stochastic gradient descent (signSGD) is a suitable candidate for large-batch ResNet training with improved generalization compared to vanilla SGD. We further modify signSGD to improve the convergence for large batches while retaining its generalization properties. In particular, we propose r-signSGD, which restricts certain parameter updates of signSGD—r-signSGD updates a parameter of the neural network only if its gradient sign matches its gradient momentum sign. This prevents unnecessary/wrong …