Lubna Farhi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Lubna Farhi has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
Farhan Ur Rehman
Department of Mechanical Engineering, University of Toronto, Toronto, Canada
3 shared publicationsHassan Imam
Department of Electronic Engineering, Sir Syed University of Engineering and Technology, Karachi, Pakistan
2 shared publications- Abdeldime Mohamed Salih Abdelgader
School of Information Science and Engineering, Southeast University, Nanjing 210096, China
1 shared publication Agha Yasir
Sir Syed University of Engineering and Technology
1 shared publication- Agha Yasir Ali
Department of Applied Computing and Emerging Sciences, Sir Syed University of Engineering and Technology, Karachi, Pakistan
1 shared publication - Agha Yasir Ali
Sir Syed University of Engineering and Technology Karachi, Pakistan
1 shared publication
Latest Publications
- Open Access
- 1 November 2022
- Published by VFAST Research Platform in VFAST Transactions on Software Engineering
- Vol. 10 (4) , 168-174
Hybrid technique to predict the position of Vehicle
Utilizing wireless networking to estimate the position of vehicles on the road is an important aspect of vehicular safety. It can be used to predetermine on-coming collisions and warn the drivers with ample time to either prevent the crash or minimize damage. The proposed methodology takes advantage of multipath effect by using the hypothesis test to detect a disturbance in the vehicle position estimate. The benefit of this technique is that it reduces position estimate error by integrating all of the vehicle’s data with Time Difference of Arrival (TDOA) and then passing that information through a Kalman filter to smooth the range data thereby decreasing Non-Line of Sight (NLOS) errors as well. The proposed technique accounts for error in the GPS data and reduces random error influences to significantly improve precision in position estimation and accuracy. The simulations showed that the Kalman filter improved accuracy by 5.45% and precision by 3.56% on an average as compared to TDOA position estimation without Kalman Filter.
- 1 November 2022
- Published by VFAST Research Platform in VFAST Transactions on Software Engineering
- Vol. 10 (4) , 168-174
Open AccessHybrid technique to predict the position of Vehicle
Utilizing wireless networking to estimate the position of vehicles on the road is an important aspect of vehicular safety. It can be used to predetermine on-coming collisions and warn the drivers with ample time to either prevent the crash or minimize damage. The proposed methodology takes advantage of multipath effect by using the hypothesis test to detect a disturbance in the vehicle position estimate. The benefit of this technique is that it reduces position estimate error by integrating all of the vehicle’s data with Time Difference of Arrival (TDOA) and then passing that information through a Kalman filter to smooth the range data thereby decreasing Non-Line of Sight (NLOS) errors as well. The proposed technique accounts for error in the GPS data and reduces random error influences to significantly improve precision in position estimation and accuracy. The simulations showed that the Kalman filter improved accuracy by 5.45% and precision by 3.56% on an average as compared to TDOA position estimation without Kalman Filter.
- Open Access
- 26 August 2022
- Published by Hindawi Limited in Wireless Communications and Mobile Computing
- Vol. 2022, 1-8
Capacity Optimization of MISO System in Intercell Interference of Visible Light Communication System
The capacity of multiple input and single out (MISO) system is investigated in intercell interference environment of a visible light communication (VLC) system. The interference is dominating in cell edges and the center of the room due to overlapping of light signals from neighbor cell. The footprint of one cell extends at high SNR, and consequently, this increases the intercell interference regions. Therefore, the channel capacity is greatly deteriorated due to interference power. In this paper, the channel capacity can be improved using diversity gain of MISO-based VLC system. The light-emitting diode (LEDs) of MISO system in a cell can use the same data to improve the diversity gain of the system. The MISO array is derived as per SNR requirements in interference region of the room. The capacity of the MISO system is examined and compared with that of the single input and single output (SISO). A trade-off of the MISO array is obtained at different SNRs to achieve the interference free capacity system. Theoretical and simulation capacity response show that the interference is considerably minimized by using diversity gain of MISO system. The interference free capacity is achieved at 44dB, 49.5dB, and 51.7dB SNR using 4X1, 9X1 and 16X MISO system, respectively.
- 26 August 2022
- Published by Hindawi Limited in Wireless Communications and Mobile Computing
- Vol. 2022, 1-8
Open AccessCapacity Optimization of MISO System in Intercell Interference of Visible Light Communication System
The capacity of multiple input and single out (MISO) system is investigated in intercell interference environment of a visible light communication (VLC) system. The interference is dominating in cell edges and the center of the room due to overlapping of light signals from neighbor cell. The footprint of one cell extends at high SNR, and consequently, this increases the intercell interference regions. Therefore, the channel capacity is greatly deteriorated due to interference power. In this paper, the channel capacity can be improved using diversity gain of MISO-based VLC system. The light-emitting diode (LEDs) of MISO system in a cell can use the same data to improve the diversity gain of the system. The MISO array is derived as per SNR requirements in interference region of the room. The capacity of the MISO system is examined and compared with that of the single input and single output (SISO). A trade-off of the MISO array is obtained at different SNRs to achieve the interference free capacity system. Theoretical and simulation capacity response show that the interference is considerably minimized by using diversity gain of MISO system. The interference free capacity is achieved at 44dB, 49.5dB, and 51.7dB SNR using 4X1, 9X1 and 16X MISO system, respectively.
- Open Access
- 25 May 2022
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2022, 1-13
Dermoscopic Image Classification Using Deep Belief Learning Network Architecture
In this paper, deep belief learning network architecture (DBL) is proposed for medical image classification in a bid to improve the diagnostics of dermal melanoma as an alternative to traditional dermoscopy. Preprocessing was carried out by using a linear Gaussian filter by eliminating high-frequency artifacts and distortion. The -means segmentation technique was used to extract the region of interest. The DBL network was then applied to the segmented image for classification. The DBL architecture disperses the weights and hyperparameters to all positions in an image, making it possible to scale to various image sizes. The effects of overfitting were mitigated for small datasets and were achieved by optimizing the proposed network. The algorithm works effectively by fine-tuning constraints. The results showed an increase in the accuracy between the proposed model and AlexNet and LeeNet for segmented images from 8% to 47%, respectively. Similarly, an increase for nonsegmented images was observed between 2% and 48%. An average reduction of 47.8% and 41.5% in error for both segmented and nonsegmented images was recorded for dermal images. The execution time also decreased in comparison with the other architectures averaged by 8-13%, since the weights were distributed only on the clustered regions in the segmented image, as compared to the whole image thus allowing the network to classify it faster with improved accuracy.
- 25 May 2022
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2022, 1-13
Open AccessDermoscopic Image Classification Using Deep Belief Learning Network Architecture
In this paper, deep belief learning network architecture (DBL) is proposed for medical image classification in a bid to improve the diagnostics of dermal melanoma as an alternative to traditional dermoscopy. Preprocessing was carried out by using a linear Gaussian filter by eliminating high-frequency artifacts and distortion. The -means segmentation technique was used to extract the region of interest. The DBL network was then applied to the segmented image for classification. The DBL architecture disperses the weights and hyperparameters to all positions in an image, making it possible to scale to various image sizes. The effects of overfitting were mitigated for small datasets and were achieved by optimizing the proposed network. The algorithm works effectively by fine-tuning constraints. The results showed an increase in the accuracy between the proposed model and AlexNet and LeeNet for segmented images from 8% to 47%, respectively. Similarly, an increase for nonsegmented images was observed between 2% and 48%. An average reduction of 47.8% and 41.5% in error for both segmented and nonsegmented images was recorded for dermal images. The execution time also decreased in comparison with the other architectures averaged by 8-13%, since the weights were distributed only on the clustered regions in the segmented image, as compared to the whole image thus allowing the network to classify it faster with improved accuracy.
- Open Access
- 13 May 2022
- Published by VFAST Research Platform in VFAST Transactions on Software Engineering
- Vol. 10 (2) , 101-108
Energy Based Segmentation for Lungs Surrounding Pulmonic Diffusion in CT Images.
Increasing number of COVID-19 Positive cases has lead for an immediate analytical requirement and fast decision making method that uses an adaptive algorithm to formulate an energy based segmentation technique. The proposed algorithm is developed using Level-Set to achieve the required results. This hybrid type method will use local along with global energies that happens to be very efficient for matching the patterns, segmenting the selected area and depends on tracing the anatomic type structures through exploiting the constraints that can be extracted out of the data set containing CT Images. This framework can perform an independent stochastic segmentation of COVID-19 in CT Imaging by smartly combining the level-set, region based, global along with three different types of energies being established as uniform modling energy (UE), mean separation energy (ME) and histogram separation energy (HE) all under same archetecture. The selection of Level-Set technique has solved the topology of the problem by defining the segmentation of boundary for local region that consist of statistics global in nature and energies local in nature given at individual point. Then strategically updating the local region by altering the energies either minimizing or maximizing it as desired. Furthermore, the proposed framework is analysed using different CT Images. The results obtained from this analysis shows that suggested framework achieves 89.5% similarity between segmented and ground truth by dice method, and by Hausdorff algorithm a minimum distance of 0.5(mm). The adaptive stochastic segmentation method in proposed framework can be applied to segment out different levels when there binary thresholds levels are greater than 0.3. The algorithm model is so designed that it can segment out the COVID-19 effected regions automatically from raw CT images with higher accurate boundaries that relatively works. The feature of this stochastic segmentation is associated with COVID-19 severity that arbitrate the causal of the COVID-19 severity.
- 13 May 2022
- Published by VFAST Research Platform in VFAST Transactions on Software Engineering
- Vol. 10 (2) , 101-108
Open AccessEnergy Based Segmentation for Lungs Surrounding Pulmonic Diffusion in CT Images.
Increasing number of COVID-19 Positive cases has lead for an immediate analytical requirement and fast decision making method that uses an adaptive algorithm to formulate an energy based segmentation technique. The proposed algorithm is developed using Level-Set to achieve the required results. This hybrid type method will use local along with global energies that happens to be very efficient for matching the patterns, segmenting the selected area and depends on tracing the anatomic type structures through exploiting the constraints that can be extracted out of the data set containing CT Images. This framework can perform an independent stochastic segmentation of COVID-19 in CT Imaging by smartly combining the level-set, region based, global along with three different types of energies being established as uniform modling energy (UE), mean separation energy (ME) and histogram separation energy (HE) all under same archetecture. The selection of Level-Set technique has solved the topology of the problem by defining the segmentation of boundary for local region that consist of statistics global in nature and energies local in nature given at individual point. Then strategically updating the local region by altering the energies either minimizing or maximizing it as desired. Furthermore, the proposed framework is analysed using different CT Images. The results obtained from this analysis shows that suggested framework achieves 89.5% similarity between segmented and ground truth by dice method, and by Hausdorff algorithm a minimum distance of 0.5(mm). The adaptive stochastic segmentation method in proposed framework can be applied to segment out different levels when there binary thresholds levels are greater than 0.3. The algorithm model is so designed that it can segment out the COVID-19 effected regions automatically from raw CT images with higher accurate boundaries that relatively works. The feature of this stochastic segmentation is associated with COVID-19 severity that arbitrate the causal of the COVID-19 severity.
- 14 February 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Autonomous Car based on Neural Network
Autonomous vehicles will frame the foundation of future cutting edge in transportation giving travel relaxation, road protection, alongside various worth added services. Such a change which will be fueled by boosts in technologies for machine learning and wireless communication will empower a future vehicular biological system that is better highlighted and more proficient. These days, endeavors are being made to give insignificant control to man over the vehicle. The point is to decrease fender benders and give individuals an agreeable climate which is the justification for turning into an autonomous vehicle. The benefits of autonomous vehicles, like less car accidents, expanded dependability, expanded street limit, diminished gridlock just as decrease of traffic police and care protection, despite the fact that we need to beat the issues of network safety, software dependability, and loss of driver related jobs. Our system is focused on a vehicle that is agreeable and affordable for people. Our autonomous car will navigate itself through the track with the help of computer vision technique. User will be capable to observe the stay view from the webcam so that it will be connected to the car. Lane detection algorithm is used to provide the required control to the car that is built through neural networks. People are free to concentrate on other work or to relaxation all through their trips.
- 14 February 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Autonomous Car based on Neural Network
Autonomous vehicles will frame the foundation of future cutting edge in transportation giving travel relaxation, road protection, alongside various worth added services. Such a change which will be fueled by boosts in technologies for machine learning and wireless communication will empower a future vehicular biological system that is better highlighted and more proficient. These days, endeavors are being made to give insignificant control to man over the vehicle. The point is to decrease fender benders and give individuals an agreeable climate which is the justification for turning into an autonomous vehicle. The benefits of autonomous vehicles, like less car accidents, expanded dependability, expanded street limit, diminished gridlock just as decrease of traffic police and care protection, despite the fact that we need to beat the issues of network safety, software dependability, and loss of driver related jobs. Our system is focused on a vehicle that is agreeable and affordable for people. Our autonomous car will navigate itself through the track with the help of computer vision technique. User will be capable to observe the stay view from the webcam so that it will be connected to the car. Lane detection algorithm is used to provide the required control to the car that is built through neural networks. People are free to concentrate on other work or to relaxation all through their trips.
- 26 January 2022
- Published by Springer Nature in Wireless Personal Communications
- Vol. 124 (2) , 1151-1167
Optimized Intelligent Auto-Regressive Neural Network Model (ARNN) for Prediction of Non-Linear Exogenous Signals
- 26 January 2022
- Published by Springer Nature in Wireless Personal Communications
- Vol. 124 (2) , 1151-1167
Optimized Intelligent Auto-Regressive Neural Network Model (ARNN) for Prediction of Non-Linear Exogenous Signals
- Open Access
- 21 December 2021
- Published by Hindawi Limited in Security and Communication Networks
- Vol. 2021, 1-11
Smart Identity Management System by Face Detection Using Multitasking Convolution Network
Identity management system in most academic and office environments is presently achieved primarily by a manual method where the user has to input their attendance into the system. The manual method sometimes results in human error and makes the process less efficient and time-consuming. The proposed system highlights the implementation and design of a smart face identification-based management system while taking into account both the background luminosity and distance. This system detects and recognizes the person and marks their attendance with the timestamp. In this methodology, the face is initially resized to 3 different sizes of 256, 384, and 512 pixels for multiscale testing. The overall outcome size descriptor is the overall mean for these characteristic vectors, and the deep convolution neural network calculates 22 facial features in 128 distinct embeddings in 22-deep network layers. The pose of the 2D face from 15 to +15° provides identification with 98 accuracy in low computation time. Another feature of the proposed system is that it is able to accurately perform identification with an accuracy of 99.92 from a distance of 5m under optimal light conditions. The accuracy is also dependent on the light intensity where it varies from 96 to 99 under 100 to 1000lumen/m2, respectively. The presented model not only improves accuracy and identity under realistic conditions but also reduces computation time.
- 21 December 2021
- Published by Hindawi Limited in Security and Communication Networks
- Vol. 2021, 1-11
Open AccessSmart Identity Management System by Face Detection Using Multitasking Convolution Network
Identity management system in most academic and office environments is presently achieved primarily by a manual method where the user has to input their attendance into the system. The manual method sometimes results in human error and makes the process less efficient and time-consuming. The proposed system highlights the implementation and design of a smart face identification-based management system while taking into account both the background luminosity and distance. This system detects and recognizes the person and marks their attendance with the timestamp. In this methodology, the face is initially resized to 3 different sizes of 256, 384, and 512 pixels for multiscale testing. The overall outcome size descriptor is the overall mean for these characteristic vectors, and the deep convolution neural network calculates 22 facial features in 128 distinct embeddings in 22-deep network layers. The pose of the 2D face from 15 to +15° provides identification with 98 accuracy in low computation time. Another feature of the proposed system is that it is able to accurately perform identification with an accuracy of 99.92 from a distance of 5m under optimal light conditions. The accuracy is also dependent on the light intensity where it varies from 96 to 99 under 100 to 1000lumen/m2, respectively. The presented model not only improves accuracy and identity under realistic conditions but also reduces computation time.
- Open Access
- 1 October 2021
- Published by Mehran University of Engineering and Technology in Mehran University Research Journal of Engineering and Technology
- Vol. 40 (4) , 747-752
Active Compliant PID Learning Control for Grinding Robots
This article proposes a Proportional, Integral, and Derivative (PID) learning controller for rigid robotic disk grinding mechanism. It has been observed that the stiffness of the robotic arm for a grinder has a direct correlation with the sensitivity of the grinding forces. It is also drastically influenced by the end-effector path tracking error resulting in limited accuracy of the robot. The error in robot’s accuracy is also increased by external interferences, such as surface imperfections and voids in the subject material. These errors can be mitigated via efficient feedback. In the proposed methodology, the controller gain is tuned by implementing a learning-based methodology to PID controllers. The learning control for the robotic grinding system helps by progressively decreasing error between the actual grinded paths and required trace. Experimental results demonstrate that as the grinder machines the required path iteratively, its grinding accuracy improves due to the learning algorithm.
- 1 October 2021
- Published by Mehran University of Engineering and Technology in Mehran University Research Journal of Engineering and Technology
- Vol. 40 (4) , 747-752
Open AccessActive Compliant PID Learning Control for Grinding Robots
This article proposes a Proportional, Integral, and Derivative (PID) learning controller for rigid robotic disk grinding mechanism. It has been observed that the stiffness of the robotic arm for a grinder has a direct correlation with the sensitivity of the grinding forces. It is also drastically influenced by the end-effector path tracking error resulting in limited accuracy of the robot. The error in robot’s accuracy is also increased by external interferences, such as surface imperfections and voids in the subject material. These errors can be mitigated via efficient feedback. In the proposed methodology, the controller gain is tuned by implementing a learning-based methodology to PID controllers. The learning control for the robotic grinding system helps by progressively decreasing error between the actual grinded paths and required trace. Experimental results demonstrate that as the grinder machines the required path iteratively, its grinding accuracy improves due to the learning algorithm.
- Open Access
- 16 March 2021
- Published by Springer Nature in Research Square
Optimized Intelligent Auto-Regressive Neural Network Model (ARNN) for Prediction of Non-Linear Exogenous Signals
The paper presents a prediction of non-linear exogenous signal by optimized intelligent auto-regressive neural network model (ARNN). A signal comprises of two sets of data called deterministic and error. The former type of data represents the degradation index of a signal, while the error is the uncertainties associated with the signal. To understand and predict signals, a intelligent approach is taken through the use of ARNN model. In this approach, the rst step is to diagnose whether a time series signal is normally distributed or not by utilizing the Jarque-Bera test. The high and low volatility data ele- ments can be separated via kurtosis hypothesis. The deterministic component of the signal is also predicted by developing a neural network based non-linear autoregressive model (NN-NARX) and the error component by using a linear model. The nal forecast is formed by combining the results determined from each of the models and evaluated using the mean square error results. Vali- dation of the prediction is obtained through a comparison of the results with other models such as ARNN, traditional ARMX, and NARX models. The re- sults show that the proposed model provides improved predictions, minimize high dependence on design parameters with low computational cost.
- 16 March 2021
- Published by Springer Nature in Research Square
Open AccessOptimized Intelligent Auto-Regressive Neural Network Model (ARNN) for Prediction of Non-Linear Exogenous Signals
The paper presents a prediction of non-linear exogenous signal by optimized intelligent auto-regressive neural network model (ARNN). A signal comprises of two sets of data called deterministic and error. The former type of data represents the degradation index of a signal, while the error is the uncertainties associated with the signal. To understand and predict signals, a intelligent approach is taken through the use of ARNN model. In this approach, the rst step is to diagnose whether a time series signal is normally distributed or not by utilizing the Jarque-Bera test. The high and low volatility data ele- ments can be separated via kurtosis hypothesis. The deterministic component of the signal is also predicted by developing a neural network based non-linear autoregressive model (NN-NARX) and the error component by using a linear model. The nal forecast is formed by combining the results determined from each of the models and evaluated using the mean square error results. Vali- dation of the prediction is obtained through a comparison of the results with other models such as ARNN, traditional ARMX, and NARX models. The re- sults show that the proposed model provides improved predictions, minimize high dependence on design parameters with low computational cost.