arXiv subjects

Rashid Ansari

Publications and source records attributed to Rashid Ansari.

14 recordsLinked to original sources

arXiv2022

Classification of the Cervical Vertebrae Maturation (CVM) stages Using the Tripod Network

We present a novel deep learning method for fully automated detection and classification of the Cervical Vertebrae Maturation (CVM) stages. The deep convolutional neural network consists of three parallel networks (TriPodNet) independently trained with different initialization parameters. They also have a built-in set of novel directional filters that highlight the Cervical Verte edges in X-ray images. Outputs of the three parallel networks are combined using a fully connected layer. 1018 cephalometric radiographs were labeled, divided by gender, and classified according to the CVM stages. Resulting images, using different training techniques and patches, were used to train TripodNet together with a set of tunable directional edge enhancers. Data augmentation is implemented to avoid overfitting. TripodNet achieves the state-of-the-art accuracy of 81.18\% in female patients and 75.32\% in male patients. The proposed TripodNet achieves a higher accuracy in our dataset than the Swin Transformers and the previous network models that we investigated for CVM stage estimation.

eess.IV↗

arXiv2020

EEG Classification by factoring in Sensor Configuration

Electroencephalography (EEG) serves as an effective diagnostic tool for mental disorders and neurological abnormalities. Enhanced analysis and classification of EEG signals can help improve detection performance. A new approach is examined here for enhancing EEG classification performance by leveraging knowledge of spatial layout of EEG sensors. Performance of two classification models - model 1 that ignores the sensor layout and model 2 that factors it in - is investigated and found to achieve consistently higher detection accuracy. The analysis is based on the information content of these signals represented in two different ways: concatenation of the channels of the frequency bands and an image-like 2D representation of the EEG channel locations. Performance of these models is examined on two tasks, social anxiety disorder (SAD) detection, and emotion recognition using a dataset for emotion analysis using physiological signals (DEAP). We hypothesized that model 2 will significantly outperform model 1 and this was validated in our results as model 2 yielded $5$--$8\%$ higher accuracy in all machine learning algorithms investigated. Convolutional Neural Networks (CNN) provided the best performance far exceeding that of Support Vector Machine (SVM) and k-Nearest Neighbors (kNNs) algorithms.

eess.SP↗

arXiv2019

Deep Layered LMS Predictor

In this study, we present a new approach to design a Least Mean Squares (LMS) predictor. This approach exploits the concept of deep neural networks and their supremacy in terms of performance and accuracy. The new LMS predictor is implemented as a deep neural network using multiple non linear LMS filters. The network consists of multiple layers with nonlinear activation functions, where each neuron in the hidden layers corresponds to a certain FIR filter output which goes through nonlinearity. The output of the last layer is the prediction. We hypothesize that this approach will outperform the traditional adaptive filters.

eess.SP↗

arXiv2018

EEG Classification based on Image Configuration in Social Anxiety Disorder

The problem of detecting the presence of Social Anxiety Disorder (SAD) using Electroencephalography (EEG) for classification has seen limited study and is addressed with a new approach that seeks to exploit the knowledge of EEG sensor spatial configuration. Two classification models, one which ignores the configuration (model 1) and one that exploits it with different interpolation methods (model 2), are studied. Performance of these two models is examined for analyzing 34 EEG data channels each consisting of five frequency bands and further decomposed with a filter bank. The data are collected from 64 subjects consisting of healthy controls and patients with SAD. Validity of our hypothesis that model 2 will significantly outperform model 1 is borne out in the results, with accuracy $6$--$7\%$ higher for model 2 for each machine learning algorithm we investigated. Convolutional Neural Networks (CNN) were found to provide much better performance than SVM and kNNs.

cs.LG↗

arXiv2018

Automated Pain Detection from Facial Expressions using FACS: A Review

Facial pain expression is an important modality for assessing pain, especially when the patient's verbal ability to communicate is impaired. The facial muscle-based action units (AUs), which are defined by the Facial Action Coding System (FACS), have been widely studied and are highly reliable as a method for detecting facial expressions (FE) including valid detection of pain. Unfortunately, FACS coding by humans is a very time-consuming task that makes its clinical use prohibitive. Significant progress on automated facial expression recognition (AFER) has led to its numerous successful applications in FACS-based affective computing problems. However, only a handful of studies have been reported on automated pain detection (APD), and its application in clinical settings is still far from a reality. In this paper, we review the progress in research that has contributed to automated pain detection, with focus on 1) the framework-level similarity between spontaneous AFER and APD problems; 2) the evolution of system design including the recent development of deep learning methods; 3) the strategies and considerations in developing a FACS-based pain detection framework from existing research; and 4) introduction of the most relevant databases that are available for AFER and APD studies. We attempt to present key considerations in extending a general AFER framework to an APD framework in clinical settings. In addition, the performance metrics are also highlighted in evaluating an AFER or an APD system.

cs.CV↗

arXiv2018

Learning Pain from Action Unit Combinations: A Weakly Supervised Approach via Multiple Instance Learning

Patient pain can be detected highly reliably from facial expressions using a set of facial muscle-based action units (AUs) defined by the Facial Action Coding System (FACS). A key characteristic of facial expression of pain is the simultaneous occurrence of pain-related AU combinations, whose automated detection would be highly beneficial for efficient and practical pain monitoring. Existing general Automated Facial Expression Recognition (AFER) systems prove inadequate when applied specifically for detecting pain as they either focus on detecting individual pain-related AUs but not on combinations or they seek to bypass AU detection by training a binary pain classifier directly on pain intensity data but are limited by lack of enough labeled data for satisfactory training. In this paper, we propose a new approach that mimics the strategy of human coders of decoupling pain detection into two consecutive tasks: one performed at the individual video-frame level and the other at video-sequence level. Using state-of-the-art AFER tools to detect single AUs at the frame level, we propose two novel data structures to encode AU combinations from single AU scores. Two weakly supervised learning frameworks namely multiple instance learning (MIL) and multiple clustered instance learning (MCIL) are employed corresponding to each data structure to learn pain from video sequences. Experimental results show an 87% pain recognition accuracy with 0.94 AUC (Area Under Curve) on the UNBC-McMaster Shoulder Pain Expression dataset. Tests on long videos in a lung cancer patient video dataset demonstrates the potential value of the proposed system for pain monitoring in clinical settings.

cs.LG↗

arXiv2017

Device-to-Device Networking Meets Cellular via Network Coding

Utilizing device-to-device (D2D) connections among mobile devices is promising to meet the increasing throughput demand over cellular links. In particular, when mobile devices are in close proximity of each other and are interested in the same content, D2D connections such as Wi-Fi Direct can be opportunistically used to construct a cooperative (and jointly operating) cellular and D2D networking system. However, it is crucial to understand, quantify, and exploit the potential of network coding for cooperating mobile devices in the joint cellular and D2D setup. In this paper, we consider this problem, and (i) develop a network coding framework, namely NCMI, for cooperative mobile devices in the joint cellular and D2D setup, where cellular and D2D link capacities are the same, and (ii) characterize the performance of the proposed network coding framework, where we use packet completion time, which is the number of transmission slots to recover all packets, as a performance metric. We demonstrate the benefits of our network coding framework through simulations.

cs.NI↗

arXiv2017

Predicting Hospital Re-admissions from Nursing Care Data of Hospitalized Patients

Readmission rates in the hospitals are increasingly being used as a benchmark to determine the quality of healthcare delivery to hospitalized patients. Around three-fourths of all hospital re-admissions can be avoided, saving billions of dollars. Many hospitals have now deployed electronic health record (EHR) systems that can be used to study issues that trigger readmission. However, most of the EHRs are high dimen-sional and sparsely populated, and analyzing such data sets is a Big Data challenge. The effect of some of the well-known dimension reduction techniques is minimized due to presence of non-linear variables. We use association mining as a dimension reduction method and the results are used to develop models, using data from an existing nursing EHR system, for predicting risk of re-admission to the hospitals. These models can help in determining effective treatments for patients to minimize the possibility of re-admission, bringing down the cost and increasing the quality of care provided to the patients. Results from the models show significantly accurate predictions of patient re-admission.

cs.CY↗

arXiv2016

Content-Aware Network Coding over Device-to-Device Networks

Consider a scenario of broadcasting a common content to a group of cooperating mobile devices that are within proximity of each other. Devices in this group may receive partial content from the source due to packet losses over wireless broadcast links. We further consider that packet losses are different for different devices. The remaining missing content at each device can then be recovered, thanks to cooperation among the devices by exploiting device-to-device (D2D) connections. In this context, the minimum amount of time that can guarantee a complete acquisition of the common content at every device is referred to as the "completion time". It has been shown that instantly decodable network coding (IDNC) reduces the completion time as compared to no network coding in this scenario. Yet, for applications such as video streaming, not all packets have the same importance and not all devices are interested in the same quality of content. This problem is even more interesting when additional, but realistic constraints, such as strict deadline, bandwidth, or limited energy are added in the problem formulation. We assert that direct application of IDNC in such a scenario yields poor performance in terms of content quality and completion time. In this paper, we propose a novel Content and Loss-Aware IDNC scheme that improves content quality and network coding opportunities jointly by taking into account importance of each packet towards the desired quality of service (QoS) as well as the channel losses over D2D links. Our proposed Content and Loss-Aware IDNC (i) maximizes the quality under the completion time constraint, and (ii) minimizes the completion time under the quality constraint. We demonstrate the benefits of Content and Loss-Aware IDNC through simulations.

cs.IT↗

arXiv2015

Efficient refinement of GPS-based localization in urban areas using visual information and sensor parameter

An efficient method is proposed for refining GPS-acquired location coordinates in urban areas using camera images, Google Street View (GSV) and sensor parameters. The main goal is to compensate for GPS location imprecision in dense area of cities due to proximity to walls and buildings. Avail-able methods for better localization often use visual information by using query images acquired with camera-equipped mobile devices and applying image retrieval techniques to find the closest match in a GPS-referenced image data set. The search areas required for reliable search are about 1-2 sq. Km and the accuracy is typically 25-100 meters. Here we describe a method based on image retrieval where a reliable search can be confined to areas of 0.01 sq. Km and the accuracy in our experiments is less than 10 meters. To test our procedure we created a database by acquiring all Google Street View images close to what is seen by a pedestrian in a large region of downtown Chicago and saved all coordinates and orientation data to be used for confining our search region. Prior knowledge from approximate position of query image is leveraged to address complexity and accuracy issues of our search in a large scale geo-tagged data set. One key aspect that differentiates our work is that it utilizes the sensor information of GPS SOS and the camera orientation in improving localization. Finally we demonstrate retrieval-based technique are less accurate in sparse open areas compared with purely GPS measurement. The effectiveness of our approach is discussed in detail and experimental results show improved performance when compared with regular approaches.

cs.CV↗

arXiv2015

Network Coding for Cooperative Mobile Devices with Multiple Interfaces

Cooperation among mobile devices and utilizing multiple interfaces such as cellular and local area links simultaneously are promising to meet the increasing throughput demand over cellular links. In particular, when mobile devices are in the close proximity of each other and are interested in the same content, device-to-device connections such as WiFi-Direct, in addition to cellular links, can be utilized to construct a cooperative system. However, it is crucial to understand the potential of network coding for cooperating mobile devices with multiple interfaces. In this paper, we consider this problem, and (i) develop network coding schemes for cooperative mobile devices with multiple interfaces, and (ii) characterize the performance of network coding by using the number of transmissions to recover all packets as a performance metric.

cs.IT↗

arXiv2013

Adaptive Hierarchical Data Aggregation using Compressive Sensing (A-HDACS) for Non-smooth Data Field

Compressive Sensing (CS) has been applied successfully in a wide variety of applications in recent years, including photography, shortwave infrared cameras, optical system research, facial recognition, MRI, etc. In wireless sensor networks (WSNs), significant research work has been pursued to investigate the use of CS to reduce the amount of data communicated, particularly in data aggregation applications and thereby improving energy efficiency. However, most of the previous work in WSN has used CS under the assumption that data field is smooth with negligible white Gaussian noise. In these schemes signal sparsity is estimated globally based on the entire data field, which is then used to determine the CS parameters. In more realistic scenarios, where data field may have regional fluctuations or it is piecewise smooth, existing CS based data aggregation schemes yield poor compression efficiency. In order to take full advantage of CS in WSNs, we propose an Adaptive Hierarchical Data Aggregation using Compressive Sensing (A-HDACS) scheme. The proposed schemes dynamically chooses sparsity values based on signal variations in local regions. We prove that A-HDACS enables more sensor nodes to employ CS compared to the schemes that do not adapt to the changing field. The simulation results also demonstrate the improvement in energy efficiency as well as accurate signal recovery.

cs.NI↗

arXiv2013

Energy Efficient Decentralized Detection Based on Bit-optimal Multi-hop Transmission in One-dimensional Wireless Sensor Networks

Existing information theoretic work in decentralized detection is largely focused on parallel configuration of Wireless Sensor Networks (WSNs), where an individual hard or soft decision is computed at each sensor node and then transmitted directly to the fusion node. Such an approach is not efficient for large networks, where communication structure is likely to comprise of multiple hops. On the other hand, decentralized detection problem investigated for multi-hop networks is mainly concerned with reducing number and/or size of messages by using compression and fusion of information at intermediate nodes. In this paper an energy efficient multi-hop configuration of WSNs is proposed to solve the detection problem in large networks with two objectives: maximizing network lifetime and minimizing probability of error in the fusion node. This optimization problem is considered under the constraint of total consumed energy. The two objectives mentioned are achieved simultaneously in the multi-hop configuration by exploring tradeoffs between different path lengths and number of bits allocated to each node for quantization. Simulation results show significant improvement in the proposed multi-hop configuration compared with the parallel configuration in terms of energy efficiency and detection accuracy for different size networks, especially in larger networks.

cs.DC↗

arXiv2012

Power-efficient Hierarchical Data Aggregation using Compressive Sensing in WSN

Compressive Sensing (CS) method is a burgeoning technique being applied to diverse areas including wireless sensor networks (WSNs). In WSNs, it has been studied in the context of data gathering and aggregation, particularly aimed at reducing data transmission cost and improving power efficiency. Existing CS based data gathering work in WSNs assume fixed and uniform compression threshold across the network, regard- less of the data field characteristics. In this paper, we present a novel data aggregation architecture model that combines a multi- resolution structure with compressed sensing. The compression thresholds vary over the aggregation hierarchy, reflecting the underlying data field. Compared with previous relevant work, the proposed model shows its significant energy saving from theoretical analysis. We have also implemented the proposed CS- based data aggregation framework on a SIDnet SWANS platform, discrete event simulator commonly used for WSN simulations. Our experiments show substantial energy savings, ranging from 37% to 77% for different nodes in the networking depending on the position of hierarchy.

cs.DC↗