Papers by Syed Farzan Ali
IEEE Access
Diabetic Retinopathy (DR) is an eye disorder in patients with diabetes. Detection of DR presence ... more Diabetic Retinopathy (DR) is an eye disorder in patients with diabetes. Detection of DR presence and its complications using fundus images at an early stage helps prevent its progression to the advanced levels. In the recent years, several well-designed Convolutional Neural Networks (CNN) have been proposed to detect the presence of DR with the help of publicly available datasets. However, these existing CNN-based classifiers focus on utilizing different architectural settings to improve the performance of detection task only i.e. presence or absence of DR. The further classification of the severity and type of the disease, however, remains a non-trivial task. To this end, we propose a multi-stream ensemble deep network to classify diabetic retinopathy severity. The proposed approach takes advantages of the deep networks and principal component analysis (PCA) to learn inter-class and intra-class variations from the raw image features. Ensemble machine learning classifiers are then applied to achieve high classification accuracy and robust performance on the obtained deep features. Specifically, a multi-stream network is made using pretrained deep learning architectures i.e. ResNet-50 and DenseNet-121 to serve as the main feature extractors. Further application of PCA reduces the dimensionality of features and effectively separates the variation space of inter-class and intra-class images. Finally, an ensemble machine learning classifier using AdaBoost and random forest algorithms is built to further improve classification accuracy. The proposed approach has been compared with multiple conventional CNN-based approaches on Messidor-2 (two categories) and EyePACS (two, five categories) datasets. The experiment results show that our proposed approach achieves superior performance (upto 95.58% accuracy) and can be considered a promising method for automatic diabetic retinopathy detection. INDEX TERMS Deep learning, ResNet, random forest, diabetic retinopathy, Messidor-2, EyePACS. HAMZA MUSTAFA received the B.S. degree in information technology from the University of the Punjab and the M.S. degree in computer science from the University of Management and Technology (UMT), Lahore, Pakistan. He is a Graduate Research Scholar with the School of Systems and Technology, UMT. He is currently working as a Software Engineer at a private firm. His research interests include machine learning, computer vision, and deep learning. SYED FAROOQ ALI received the M.S. degree (Hons.) in CS from LUMS, Lahore, Pakistan and the Ph.D. degree in CS from UMT, Pakistan. He did his Ph.D. course work, the Ph.D. Comprehensive exam and the M.S. degree in CS from Ohio State University, Columbus, USA. He received the LUMS Fellowship for M.S. degree. He is currently working as an Assistant Professor with UMT. His research interests include computer vision, digital image processing, and medical imaging. He is a reviewer for various IEEE conferences and journals. MUHAMMAD BILAL is an Educator, a Researcher, and a Maker. He was a Postdoctoral Researcher at KAIST, South Korea. He is an Associate Professor with the Department of Electrical and Computer Engineering, KAU. His research interests include digital image/signal processing, machine learning/AI, digital/analog circuit design, embedded systems, and robotics.
IEEE Access
An accurate electrical Short-term Load Forecasting (STLF) is an eminent factor in the power gener... more An accurate electrical Short-term Load Forecasting (STLF) is an eminent factor in the power generation, electrical load dispatching and energy planning for the power supply companies, specifically in developing countries. This paper proposes a novel temporal feature selection-based Long Short-term Memory (LSTM) model developed by the combination of standard Artificial Neural Network (ANN) layer and LSTM for electrical short term load forecasting. The LSTM model has excellent capability of predicting the stochastic nature of an hour ahead electrical loads. The standard ANN layer consisting 11 neurons is used as an input to LSTM cells. Such a combination of ANN layer with LSTM was never proposed before. The proposed model accommodates variations in weather as well as temporal inputs like humidity, holidays, and date-time features in the hourly load data of the power supply company situated in Johor, Malaysia. This paper gives the insights of hyper parameter tuning to capture the more generalized electrical load patterns in the dataset without compromising the time complexity of the proposed model. The proposed approach was compared with five existing approaches, namely: ANN, LSTM model 1, LSTM model 2, LSTM model 3 and Convolutional Neural Network-LSTM (CNN-LSTM) using hourly load dataset of Johor. The experimental results demonstrate that the proposed approach outperformed the existing approaches in terms of root mean square error, mean absolute percentage error and Diebold-Mariano statistical inference test within 95% confidence interval. INDEX TERMS Artificial neural network, deep learning, load forecasting, long short-term memory.
Multimedia Tools and Applications
Applied Sciences
In the last decade, distraction detection of a driver gained a lot of significance due to increas... more In the last decade, distraction detection of a driver gained a lot of significance due to increases in the number of accidents. Many solutions, such as feature based, statistical, holistic, etc., have been proposed to solve this problem. With the advent of high processing power at cheaper costs, deep learning-based driver distraction detection techniques have shown promising results. The study proposes ReSVM, an approach combining deep features of ResNet-50 with the SVM classifier, for distraction detection of a driver. ReSVM is compared with six state-of-the-art approaches on four datasets, namely: State Farm Distracted Driver Detection, Boston University, DrivFace, and FT-UMT. Experiments demonstrate that ReSVM outperforms the existing approaches and achieves a classification accuracy as high as 95.5%. The study also compares ReSVM with its variants on the aforementioned datasets.
Applied Sciences, 2021
Over the last decade, a driver’s distraction has gained popularity due to its increased significa... more Over the last decade, a driver’s distraction has gained popularity due to its increased significance and high impact on road accidents. Various factors, such as mood disorder, anxiety, nervousness, illness, loud music, and driver’s head rotation, contribute significantly to causing a distraction. Many solutions have been proposed to address this problem; however, various aspects of it are still unresolved. The study proposes novel geometric and spatial scale-invariant features under a boosting framework for detecting a driver’s distraction due to the driver’s head panning. These features are calculated using facial landmark detection algorithms, including the Active Shape Model (ASM) and Boosted Regression with Markov Networks (BoRMaN). The proposed approach is compared with six existing state-of-the-art approaches using four benchmark datasets, including DrivFace dataset, Boston University (BU) dataset, FT-UMT dataset, and Pointing’04 dataset. The proposed approach outperforms the ...
Sensors, 2021
Diabetic retinopathy, an eye disease commonly afflicting diabetic patients, can result in loss of... more Diabetic retinopathy, an eye disease commonly afflicting diabetic patients, can result in loss of vision if prompt detection and treatment are not done in the early stages. Once the symptoms are identified, the severity level of the disease needs to be classified for prescribing the right medicine. This study proposes a deep learning-based approach, for the classification and grading of diabetic retinopathy images. The proposed approach uses the feature map of ResNet-50 and passes it to Random Forest for classification. The proposed approach is compared with five state-of-the-art approaches using two category Messidor-2 and five category EyePACS datasets. These two categories on the Messidor-2 dataset include ’No Referable Diabetic Macular Edema Grade (DME)’ and ’Referable DME’ while five categories consist of ‘Proliferative diabetic retinopathy’, ‘Severe’, ‘Moderate’, ‘Mild’, and ‘No diabetic retinopathy’. The results show that the proposed approach outperforms compared approaches ...
IEEE Access, 2021
Rice is a high valued subsistence crop that feeds more than 3.5 billion of the world population. ... more Rice is a high valued subsistence crop that feeds more than 3.5 billion of the world population. Its importance can be gauged from the fact that the top five rice exporting countries had a combined net export worth of around 19 billion dollars in 2018. A robust rice grain analysis and classification system can significantly improve performance both in terms of accuracy as well as time. In recent decades, this research area has garnered a lot of attention due to its socioeconomic impact. In this paper, we reviewed the work done in image-based rice classification and gradation. The contribution of this study is threefold. First, it divides the algorithms and techniques of this area into five different approaches namely; geometric, statistical, supervised, unsupervised, and deep learning. Among these, deep learning techniques have shown more promising results and gained attention for future research. Secondly, it divides the rice grain literature historically into three different eras. Thirdly, it summarizes various algorithms and techniques related to rice quality grading and rice disease identification.
Frontiers of Computer Science, 2020
Fingerprint matching, spoof mitigation and liveness detection are the trendiest biometric techniq... more Fingerprint matching, spoof mitigation and liveness detection are the trendiest biometric techniques, mostly because of their stability through life, uniqueness and their least risk of invasion. In recent decade, several techniques are presented to address these challenges over well-known data-sets. This study provides a comprehensive review on the fingerprint algorithms and techniques which have been published in the last few decades. It divides the research on fingerprint into nine different approaches including feature based, fuzzy logic, holistic, image enhancement, latent, conventional machine learning, deep learning, template matching and miscellaneous techniques. Among these, deep learning approach has outperformed other approaches and gained significant attention for future research. By reviewing fingerprint literature, it is historically divided into four eras based on 106 referred papers and their cumulative citations.
Energy Reports, 2021
Pakistan needs to overcome the cost of power generation and the ever-increasing demand for energy... more Pakistan needs to overcome the cost of power generation and the ever-increasing demand for energy with environment-friendly renewable energy resources. Several research efforts have been made with the support of Pakistan Meteorological Department in the last two decades for wind resource assessment (WRA) across the country. However, the practical installation of wind farms is quite a fraction of the total forecast wind energy potential. In this feasibility, WRA of Umerkot and Sujawal districts located in Sindh provinces of Pakistan has been analyzed by analyzing mean wind speeds, estimated Weibull parameters, power and energy densities calculation for various heights of selected wind turbines. Further, this paper analyzes the overall energy potential for these locations with implementation cost and pay-back period for investment. These locations are selected by the World Bank initiative of wind profiling campaigns to record wind speed data during 2016 and 2018 with 10 min resolution. It is observed that Umerkot and Sujawal sites are suitable for energy production. The highest values of power and energy densities for Sujawal are 414.18 W/m 2 and 3628.22 kWh/m 2 /Yr and for Umerkot these values are 303.86 W/m 2 and 2661.81 kWh/m 2 /Yr. The results indicate that using Nordex N90/2500 wind turbines are highly beneficial for Umerkot and Sujawal. The associated costs of energy are 0.074 $/kWh and 0.056 $kWh respectively and the payback period is estimated to be around 7 years with 20 years life time of the project. This work suggests the possibility of wind farm installation and commissioning based on power density calculation and cost of land acquisition. This work emphasizes the investment for wind farms at Sujawal and Umerkot for the sustainable growth of the country. This helps out policymakers for long term planning, development of wind energy projects and attracting investment for the country.
Artificial Intelligence Review, 2019
Computer vision systems open a new challenge to recognize human faces under varied poses in simil... more Computer vision systems open a new challenge to recognize human faces under varied poses in similar capacity and capability as human-beings perform naturally. For surveillance applications, pose-invariant face recognition (PIFR) will become a major breakthrough by presenting the solution of this unique challenge. In recent decade, several techniques are presented to address this challenge over well-known data-sets. These efforts are divided chronologically into seven different approaches say geometric, statistical, holistic, template, supervised learning, unsupervised learning and deep learning. Among these deep learning techniques have shown more promising results and have gained attention for future research. By reviewing PIFR, it is historically divided into five eras based on 160 referred papers and their cumulative citations.
International Journal of Turbomachinery, Propulsion and Power, 2019
A parametric study of a novel turbofan engine with an auxiliary high-pressure bypass (AHPB) is pr... more A parametric study of a novel turbofan engine with an auxiliary high-pressure bypass (AHPB) is presented. The underlying motivation for the study was to introduce and explore a configuration of a turbofan engine which could facilitate clean secondary burning of fuel at a higher temperature than conventionally realized. The study was also motivated by the developments in engineering materials for high-temperature applications and the potential utility of these developments. The parametric study is presented in two phases. Phase I presents a schematic of the turbofan engine with AHPB and the mathematics of the performance parameters at various stations. The proposed engine is hypothesized to consist of three streams—core stream, low-pressure bypass (LPB) stream, and the AHPB or, simply, the high-pressure bypass (HPB) stream. Phase II delves into the performance simulation and the analysis of the results in an ideal set-up. The simulation and results are presented for performance analy...
VAWKUM Transactions on Computer Sciences, 2018
Interest in research activities in facial processing especially cartoonization, caricature and em... more Interest in research activities in facial processing especially cartoonization, caricature and emotion generation have gradually increased over the recent years. The contribution of this paper is threefold. First, it provides an algorithms along with its results in which exaggerated cartoon like effects are added into a single facial frontal image according to the given cartoon template. The cartoon formed in this process will have similar features as of original image. Secondly, the study provides facial transformation algorithms and techniques to generate various artifacts and emotions including sad, shy, happy, blank, serious, surprise and innocent. Thirdly, the study discusses different transformation algorithms to generate various caricatures from the single frontal facial image. The output images generated using these open source algorithms and techniques are also provided in this paper to assess their subjective quality.
VAWKUM Transactions on Computer Sciences, 2018
E-Learning and Virtual Classroom applications have gained a lot of popularity due to growing popu... more E-Learning and Virtual Classroom applications have gained a lot of popularity due to growing population, easy access and low cost solution. The study in this paper proposes an open source Virtual Classroom application that tends to mimic all the functionalities and features of real class room. Its interface designs are based on the online learning theories. It will provide the students and teachers a real time virtual platform, where they can learn, share and properly propagate their knowledge, views and ideas. This open source application allows the faculty members to conduct all the class activities as if they are in real classroom. On the other hand, students have the advantage of raising questions during the lecture with the help of a chat box and a white board. In order to assess the interfaces of this application, Micrsoft Visual Studio 2012 has been used. Our application provides security and reliability to all its users. All the courses, students and faculty members are managed in a real time using this application. Administrator handles all these procedures, and has all the rights over the system including the users and databases.
Sensors (Basel, Switzerland), Jan 12, 2018
Fall induced damages are serious incidences for aged as well as young persons. A real-time automa... more Fall induced damages are serious incidences for aged as well as young persons. A real-time automatic and accurate fall detection system can play a vital role in timely medication care which will ultimately help to decrease the damages and complications. In this paper, we propose a fast and more accurate real-time system which can detect people falling in videos captured by surveillance cameras. Novel temporal and spatial variance-based features are proposed which comprise the discriminatory motion, geometric orientation and location of the person. These features are used along with ensemble learning strategy of boosting with J48 and Adaboost classifiers. Experiments have been conducted on publicly available standard datasets including () and achieving percentage accuracies of 99.2, 99.25 and 99.0, respectively. Comparisons with nine state-of-the-art methods demonstrate the effectiveness of the proposed approach on both datasets.
Lecture Notes in Computer Science, 2014
In this study, we focus on the analytical derivation of the expected distance between all sensor ... more In this study, we focus on the analytical derivation of the expected distance between all sensor nodes and the base station (i.e., E[dtoBS]) in a randomly deployed WSN. Although similar derivations appear in the related literature, to the best of our knowledge, our derivation, which assumes a particular scenario, has not been formulated before. In this specific scenario, the sensing field is a square-shaped region and the base station is located at some arbitrary distance to one of the edges of the square. Having the knowledge of E[dtoBS] value is important because E[dtoBS] provides a network designer with the opportunity to make a decision on whether it is energy-efficient to perform clustering for WSN applications that aim to pursue the clustered architectures. Similarly, a network designer might make use of this expected value during the process of deciding on the modes of communications (i.e., multi-hop or direct communication) after comparing it with the maximum transmission ranges of devices. Last but not least, the use of our derivation is not limited to WSN domain. It can be also exploited in any domain when there is a need for a probabilistic approach to find the average distance between any given number of points which are all assumed to be randomly and uniformly located in any square-shaped region and at a specific point outside this region.
2012 5th International Conference on New Technologies, Mobility and Security (NTMS), 2012
Clustering can be used as an effective technique to achieve both energy load balancing and an ext... more Clustering can be used as an effective technique to achieve both energy load balancing and an extended lifetime for a wireless sensor network (WSN). This paper presents a novel approach that first creates energy balanced fixed/static clusters, and then, to attain energy load balancing within each fixed cluster, rotates the role of cluster head through uniformly quantized energy levels based approach to prolong the overall network lifetime. The method provided herein, not only provides near-dynamic clustering performance but also reduces the complexity due to the fact that cluster formation phase is implemented once. The presented simulation results clearly show the efficacy of this proposed algorithm and thus, it can be used as a practical approach to obtain maximized network lifetime for energy balanced clusters in fixed clustering environments.
Studies in Computational Intelligence, 2011
The basic aim of this book chapter is to provide a survey on BBC Dirac Video Codec. That survey, ... more The basic aim of this book chapter is to provide a survey on BBC Dirac Video Codec. That survey, would not only provide the in depth description of different version of Dirac Video Codec but would also explain the algorithmic explanation of Dirac at implementation level. This chapter would not only provide help to new researchers who are working to
2009 9th International Symposium on Communications and Information Technology, 2009
This paper presents an efficient motion estimation algorithm for enhancing the performance of Dir... more This paper presents an efficient motion estimation algorithm for enhancing the performance of Dirac wavelet based video encoder. The proposed scheme is based on a 3D Recursive Search algorithm that uses the previously calculated motion vectors from the neighboring blocks ...
Sensors, 2012
Ripeness classification of oil palm fresh fruit bunches (FFBs) during harvesting is important to ... more Ripeness classification of oil palm fresh fruit bunches (FFBs) during harvesting is important to ensure that they are harvested during optimum stage for maximum oil production. This paper presents the application of color vision for automated ripeness classification of oil palm FFB. Images of oil palm FFBs of type DxP Yangambi were collected and analyzed using digital image processing techniques. Then the color features were extracted from those images and used as the inputs for Artificial Neural Network (ANN) learning. The performance of the ANN for ripeness classification of oil palm FFB was investigated using two methods: training ANN with full features and training ANN with reduced features based on the Principal Component Analysis (PCA) data reduction technique. Results showed that compared with using full features in ANN, using the ANN trained with reduced features can improve the classification accuracy by 1.66% and is more effective in developing an automated ripeness classifier for oil palm FFB. The developed ripeness classifier can act as a sensor in determining the correct oil palm FFB ripeness category.
SPE Production & Facilities, 2005
SummaryOwing to the increased cost of scale management in subsea developments, compared with plat... more SummaryOwing to the increased cost of scale management in subsea developments, compared with platform or onshore fields, and because of the more-limited opportunities for interventions, it is becoming increasingly important to carry out a risk-analysis process for scale management as early as possible in the field-development plan. This process involves identifying the potential scale risks and analyzing and comparing the options available for managing those risks.This paper discusses how this risk-analysis process should be carried out, with a strong emphasis on the need to integrate all the available production-chemistry and reservoir-engineering data. To demonstrate this process, an example is used from a development complex that lies in water depths greater than 400 m (greater than 1,300 ft) offshore west Africa. The process involves the following steps:Analysis of available brine samples to identify maximum scaling potential. Laboratory testing of available scale inhibitors to ...
Uploads
Papers by Syed Farzan Ali