Speech Emotion Recognition Using Convolution Neural Networks and Multi-Head Convolutional Transformer.

Speech emotion recognition (SER) is a challenging task in human-computer interaction (HCI) systems. One of the key challenges in speech emotion recognition is to extract the emotional features effectively from a speech utterance. Despite the promising results of recent studies, they generally do not leverage advanced fusion algorithms for the generation of effective representations of emotional features in speech utterances. To address this problem, we describe the fusion of spatial and temporal feature representations of speech emotion by parallelizing convolutional neural networks (CNNs) and a Transformer encoder for SER. We stack two parallel CNNs for spatial feature representation in parallel to a Transformer encoder for temporal feature representation, thereby simultaneously expanding the filter depth and reducing the feature map with an expressive hierarchical feature representation at a lower computational cost. We use the RAVDESS dataset to recognize eight different speech emotions. We augment and intensify the variations in the dataset to minimize model overfitting. Additive White Gaussian Noise (AWGN) is used to augment the RAVDESS dataset. With the spatial and sequential feature representations of CNNs and the Transformer, the SER model achieves 82.31% accuracy for eight emotions on a hold-out dataset. In addition, the SER system is evaluated with the IEMOCAP dataset and achieves 79.42% recognition accuracy for five emotions. Experimental results on the RAVDESS and IEMOCAP datasets show the success of the presented SER system and demonstrate an absolute performance improvement over the state-of-the-art (SOTA) models.

Low-Complexity Beamforming Design for a Cooperative Reconfigurable Intelligent Surface-Aided Cell-Free Network.

Cell-free (CF) networks are proposed to suppress the interference among collocated cells by deploying several BSs without cell boundaries. Nevertheless, as installing several base stations (BSs) may require high power consumption, cooperative CF networks integrated with a reconfigurable intelligent surface (RIS)/metasurface can avoid this problem. In such cooperative RIS-aided MIMO networks, efficient beamforming schemes are essential to boost their spectral and energy efficiency. However, most of the existing available beamforming schemes to maximize spectral and energy efficiency are complex and entail high complexity due to the matrix inversions. To this end, in this work we present a computationally efficient stochastic optimization-based particle swarm optimization (PSO) algorithm to amplify the spectral efficiency of the cooperative RIS-aided CF MIMO system. In the proposed PSO algorithm, several swarms are generated, while the direction of each swarm is tuned in each iteration based on the sum-rate performance to obtain the best solution. Our simulation results show that our proposed scheme can approximate the performance of the existing solutions for both the performance metrics, i.e., spectral and energy efficiency, at a very low complexity.

End-to-End Deep Convolutional Recurrent Models for Noise Robust Waveform Speech Enhancement.

Because of their simple design structure, end-to-end deep learning (E2E-DL) models have gained a lot of attention for speech enhancement. A number of DL models have achieved excellent results in eliminating the background noise and enhancing the quality as well as the intelligibility of noisy speech. Designing resource-efficient and compact models during real-time processing is still a key challenge. In order to enhance the accomplishment of E2E models, the sequential and local characteristics of speech signal should be efficiently taken into consideration while modeling. In this paper, we present resource-efficient and compact neural models for end-to-end noise-robust waveform-based speech enhancement. Combining the Convolutional Encode-Decoder (CED) and Recurrent Neural Networks (RNNs) in the Convolutional Recurrent Network (CRN) framework, we have aimed at different speech enhancement systems. Different noise types and speakers are used to train and test the proposed models. With LibriSpeech and the DEMAND dataset, the experiments show that the proposed models lead to improved quality and intelligibility with fewer trainable parameters, notably reduced model complexity, and inference time than existing recurrent and convolutional models. The quality and intelligibility are improved by 31.61% and 17.18% over the noisy speech. We further performed cross corpus analysis to demonstrate the generalization of the proposed E2E SE models across different speech datasets.

Node Calibration in UWB-Based RTLSs Using Multiple Simultaneous Ranging.

Ultra-wideband (UWB) networks are gaining wide acceptance in short- to medium-range wireless sensing and positioning applications in indoor environments due to their capability of providing high-ranging accuracy. However, the performance is highly related to the accuracy of measured position and antenna delay of anchor nodes, which form a reference positioning system of fixed infrastructure nodes. Usually, the position and antenna delay of the anchor nodes are measured separately as a standard initial procedure. Such separate measurement procedures require relatively more time and manual interventions. This paper presents a system that simultaneously measures the position and antenna delay of the anchor nodes. It provides comprehensive mathematical modeling, design, and implementation of the proposed system. An experimental evaluation in a line-of-sight (LOS) environment shows the effectiveness of the anchor nodes, whose position and antenna delay values are measured by the proposed system, in localizing a mobile node.

Safety of laparoscopic pancreaticoduodenectomy in patients with liver cirrhosis using propensity score matching.

Currently, safety of laparoscopic pancreaticoduodenectomy (LPD) in patients with liver cirrhosis is unknown. The aim of this study was to explore postoperative morbidity and mortality and long-term outcomes of cirrhotic patients after LPD. The study was a one-center retrospective study comprising 353 patients who underwent LPD between October 2010 and December 2019. A total of 28 patients had liver cirrhosis and were paired with 56 non-cirrhotic counterparts through propensity score matching (PSM). Baseline data, intra-operative data, postoperative data, and survival data were collected. Postoperative morbidity was considered as primary outcome whereas postoperative mortality, surgical parameters (operative durations, intraoperative blood loss), and long-term overall survival were secondary outcomes. Cirrhotic patients showed postoperative complication rates of 82% compared with rates of patients in the control group (48%) (P = 0.003). Further, Clavien-Dindo ≥III complication rates of 14% and 11% (P = 0.634), Clavien-Dindo I-II complication rates of 68% and 38% (P = 0.009), hospital mortality of 4% and 2% (P = 0.613) were observed for cirrhotic patients and non-cirrhotic patients, respectively. In addition, an overall survival rate of 32 months and 34.5 months (P = 0.991), intraoperative blood loss of 300 (200-400) ml and 150 (100-250) ml (P<0.0001), drain amount of 2572.5 (1023.8-5275) ml and 1617.5 (907.5-2700) ml (P = 0.048) were observed in the cirrhotic group and control group, respectively. In conclusion, LPD is associated with increased risk of postoperative morbidity in patients with liver cirrhosis. However, the incidence of Clavien-Dindo ≥III complications and post-operative mortality showed no significant increase. In addition, liver cirrhosis showed no correlation with poor overall survival in patients who underwent LPD. These findings imply that liver cirrhosis patients can routinely be considered for LPD at high volume centers with rigorous selection and management.

Multiple Simultaneous Ranging in IR-UWB Networks.

Growth in the applications of wireless devices and the need for seamless solutions to location-based services has motivated extensive research efforts to address wireless indoor localization networks. Existing works provide range-based localization using ultra-wideband technology, focusing on reducing the inaccuracy in range estimation due to clock offsets between different devices. This is generally achieved via signal message exchange between devices, which can lead to network congestion when the number of users is large. To address the problem of range estimation with limited signal messages, this paper proposes multiple simultaneous ranging methods based on a property of time difference of reception of two packets transmitted from different sources in impulse-radio ultra-wideband (IR-UWB) networks. The proposed method maintains similar robustness to the clock offsets while significantly reducing the air time occupancy when compared with the best existing ranging methods. Experimental evaluation of ranging in a line-of-sight environment shows that the proposed method enables accurate ranging with minimal air time occupancy.

Collaborative Resource Management for Negotiable Multi-Operator Small Cell Networks.

The ubiquitous coverage/connectivity requirement of wireless cellular networks has shifted mobile network operators' (MNOs) interest toward dense deployment of small cells with coverage areas that are much smaller as compared to macrocell base stations (MBSs). Multi-operator small cells could provide virtualization of network resources (infrastructure and spectrum) and enable its efficient utilization, i.e., uninterrupted coverage and connectivity to subscribers, and an opportunity to avoid under-utilization of the network resources. However, a MNO with exclusive ownership to network resources would have little incentive to utilize its precious resources to serve users of other MNOs, since MNOs differentiate among others based on their ownership of the licensed spectrum. Thus, considering network resources scarcity and under-utilization, this paper proposes a mechanism for multi-operator small cells collaboration through negotiation that establishes a mutual agreement acceptable to all involved parties, i.e., a win-win situation for the collaborating MNOs. It enables subscribers of a MNO to utilize other MNOs' network resources, and allows MNOs to offer small cells "as a service" to users with ubiquitous access to wireless coverage/connectivity, maximize the use of an existing network resources by serving additional users from a market share, and enhance per-user data rate. We validated and evaluated the proposed mechanism through simulations considering various performance metrics.

A rare case of central pancreatectomy for isolated complete pancreatic neck transection trauma.

BACKGROUND: Pancreatic trauma accounts for only 0.2% of blunt trauma and 1-12% of penetrating injuries. Injuries to other organs, such as spleen, liver, or kidney, are associated with 50.5% of the cases. The isolated complete traumatic transection of the pancreatic neck is rare. In the past, pancreatoduodenectomy or distal pancreatectomy and splenectomy was the standard care for patients with traumatic transection of the pancreatic head, duodenum or distal pancreas, and pancreatic neck. However, limited cases have been reported on the central pancreatectomy for pancreatic neck injuries. We present a rare case of a 21-year-old male patient who received central pancreatectomy for isolated complete traumatic transection of the pancreatic neck. CASE PRESENTATION: A 21-year-old male patient with mild abdominal pain and showed no apparent abnormality in the initial abdominal computed tomography (CT) was brought to the local hospital's emergency department due to a traffic accident. The patient's abdominal pain became progressively worse during observation in the hospital that led to the patient being referred to our hospital. The patient's vital signs were stable, and a physical examination revealed marked tenderness and rebound pain throughout the abdomen. The patient's white blood cells were increased; The serum amylase and lipase levels were elevated. The abdominal computed tomography revealed pancreatic neck parenchymal discontinuity, peripancreatic effusion, and hemorrhage. The patient underwent exploratory laparotomy. Intraoperative examination identified the neck of the pancreas was completely ruptured, and no apparent abnormalities were observed in the other organs. The patient underwent central pancreatectomy and Roux -Y pancreaticojejunostomy. The patient was treated with antibiotics, acid inhibition and nutritional supports for 10 days after surgery. Symptoms of the patient were significantly relieved, and white blood cells, serum amylase, and lipase levels returned to normal. The patient underwent follow up examination for 6 months with no evidence of exocrine or endocrine insufficiency. CONCLUSIONS: Central pancreatectomy is an effective pancreas parenchyma preserving procedure, may be a promising alternative to distal pancreatectomy and splenectomy for this complex pancreatic trauma in hemodynamically stable patients. Patient selection and surgeon experience are crucial in the technical aspects of this procedure.

Marginal Contribution-Based Distributed Subchannel Allocation in Small Cell Networks.

The paper presents a game theoretic solution for distributed subchannel allocation problem in small cell networks (SCNs) analyzed under the physical interference model. The objective is to find a distributed solution that maximizes the welfare of the SCNs, defined as the total system capacity. Although the problem can be addressed through best-response (BR) dynamics, the existence of a steady-state solution, i.e., a pure strategy Nash equilibrium (NE), cannot be guaranteed. Potential games (PGs) ensure convergence to a pure strategy NE when players rationally play according to some specified learning rules. However, such a performance guarantee comes at the expense of complete knowledge of the SCNs. To overcome such requirements, properties of PGs are exploited for scalable implementations, where we utilize the concept of marginal contribution (MC) as a tool to design learning rules of players’ utility and propose the marginal contribution-based best-response (MCBR) algorithm of low computational complexity for the distributed subchannel allocation problem. Finally, we validate and evaluate the proposed scheme through simulations for various performance metrics.