Skip to main content

S. Sethu Selvi

Followers

25

Following

3

Mentions

1

Public Views

Interests

Uploads

Papers by S. Sethu Selvi

Aircraft Parameter Estimation using Feedforward Neural Networks With Lyapunov Stability Analysis

Defence Science Journal, Nov 1, 2022

Aerodynamic parameter estimation is critical in the aviation sector, especially in design and dev... more Aerodynamic parameter estimation is critical in the aviation sector, especially in design and development programs of defense-military aircraft. In this paper, new results of the application of Artificial Neural Networks (ANN) to the field of aircraft parameter estimation are presented. The performances of Feedforward Neural Network (FFNN) with Backpropagation and FFNN with Backpropagation using Recursive Least Square (RLS) are investigated for aerodynamic parameter estimation. The methods are validated on flight data simulated using MATLAB implementations. The normalized Lyapunov energy functional has been used to derive the convergence conditions for both the ANN-based estimation algorithms. The estimation results are compared on the basis of performance metrics and computation time. The performance of FFNN-RLS has been observed to be approximately 10% better than FFNN-BPN. Simulation results from both algorithms have been found to be highly satisfactory and pave the way for further applications to real flight test data.

A Novel Deep Learning Algorithm for Covid Detection and Classification

2022 2nd International Conference on Intelligent Technologies (CONIT), Jun 24, 2022

Setnet: A Sparse Ensemble Network For Drone Localization And Zero Shot Drone Tracking In Real Time Surveillance Videos

2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)

Time Series based Models for Corona Data Analytics

2022 Fourth International Conference on Emerging Research in Electronics, Computer Science and Technology (ICERECT)

Machine Learning Based Classification of Welded Components

2022 4th International Conference on Circuits, Control, Communication and Computing (I4C)

A Deep Learning Approach to Classify Drones and Birds

2022 IEEE 2nd Mysore Sub Section International Conference (MysuruCon)

Deep Learning Approach to Nailfold Capillaroscopy Based Diabetes Mellitus Detection

International Journal of Online and Biomedical Engineering (iJOE)

Diabetes mellitus is a commonly occurring chronic metabolic disorder which has affected almost 40... more Diabetes mellitus is a commonly occurring chronic metabolic disorder which has affected almost 400 million people around the world. It can lead to vascular structure alterations and various renal, cardiovascular, and neurologicalcomplications claiming several lives. Since diabetes mellitus results is vascular structure changes, NailfoldCapillaroscopy(NFC) based approach can be employed for the detection of diabetes. NFC is an inexpensive, non-invasive method which involves acquisition of images of capillaries in the nail bed region using a USB digital microscope. Qualitative parameters of the capillaries such as tortuosity, hemorrhages, angiogenesis, elongated capillariesand quantitative parameters like length, width and mean capillary density are considered for diabetes detection. About 600 capillary images of healthy and diabetic subjects were collected and further data augmentation was performed to increase this to 1018 images dataset. This paper focuses on using NFC to obtain ca...

Effective Abnormal Behavior identification in a Crowd

2018 3rd International Conference on Circuits, Control, Communication and Computing (I4C), 2018

Effective analysis of crowd behavior is very challenging issue in video surveillance system. Most... more Effective analysis of crowd behavior is very challenging issue in video surveillance system. Most of the conventional approach is not up to the mark for real time behavior analysis. In this work, we have proposed an optical flow method along with KNCN classification approach to recognize abnormality in a video scene. Optical flow is a global motion information method which works efficiently. Optical flow vector provides direction and displacement of a point between two consecutive frames. This feature vector acts as input for KNN classifier. A surrounding neighbor (SN) concept is used for finding k-Nearest Centroid Neighbors(k-NCN) for each frame. We assign the class level based on the votes of k-NCN. The obtained result is promising and seems very efficient in terms of accuracy. We analyzed result and shown frame level comparison with ground truth. ROC analysis provides area under curve (AUC) for different data sequences and is compared with other state of the art techniques. Our proposed KNCN-OF method outperforms all other state of the art techniques with an accuracy of 99%.

Line Segmentation of Handwritten Kannada Text

Text line segmentation is an essential pre-processing stage for off-line handwriting recognition ... more Text line segmentation is an essential pre-processing stage for off-line handwriting recognition in many Optical Character Recognition (OCR) systems. It is an important step, as inaccurately segmented text lines will cause errors in the subsequent recognition stage. Text line segmentation of handwritten documents is still one of the most complicated problems in developing a reliable OCR. The nature of handwriting styles makes the process of text line segmentation very challenging. This paper provides a means of text line segmentation using Vertical Strip Projection based method. In this paper, the document is divided into vertical strips of appropriate width. Strip-wise horizontal histograms are then computed and the relationship of the peak valley points is used for line segmentation. Though the global horizontal projection method works efficiently for printed documents, it becomes highly unreliable when used in case of unconstrained handwritten text because of many factors like variable skew, touching and overlapping lines. The proposed method provides much improved and reliable results when compared to global horizontal profile method in case of skewed and variable sloping lines. Segmentation accuracy of 83% is achieved for handwritten Kannada text using the proposed method.

Multi-Algorithm Fusion for Fingerprint Recognition based on Texture Features

Establishing the identity of a person with highconfidence using biometric systems are gaining imp... more Establishing the identity of a person with highconfidence using biometric systems are gaining importance. It is achallenge to improve the recognition rate of an existing unimodalbiometric system. Fingerprint recognition is one of the most matureand proven technology because of its immutability and individuality.Recognition result of the system is based mainly on feature extractionmethod and type of matcher used. This paper proposes amulti-algorithm fusion algorithm for fingerprint recognition. The mainobjective of the proposed system is to improve performance usingtexture features. Feature extraction is based on ridge information andtextures. Orientation features, Curvelet transform features and DualTree-Complex Wavelet Transform (DT-CWT) features are extractedand using Euclidean distance match scores are evaluated. Texturefeatures need very less pre-processing compared to orientationfeatures. With this speed of the recognition system is improved.Weighted sum method is used in fusion of matchers. Performance ofindividual matchers in terms of False Acceptance Rate (FAR) andFalse Reject Rate (FRR) has been evaluated. For optimal threshold (η), percentage genuine recognition rate (%GAR) is calculated. Algorithm is tested on fingerprint database of 100 users and also with FVC2002-DB3 database. Maximum recognition rate of 95.2% is achieved by combining Curvelet and DT-CWT features

Blind Image Seperation Using Forward Difference Method (FDM)

Signal & Image Processing : An International Journal, 2011

In this paper, blind image separation is performed, exploiting the property of sparseness to repr... more In this paper, blind image separation is performed, exploiting the property of sparseness to represent images. A new sparse representation called forward difference method is proposed. It is known that most of the independent component analysis (ICA) basis functions, extracted from images are sparse and gives unreliable sparseness measure. In the proposed method, the image mixture is first transformed to sparse images. These images are divided into blocks and for each block the sparseness measure 0 norm is applied. The block having the most sparseness is considered to determine the separation matrix. The efficiency of the proposed method is compared with other sparse representation functions.

A Novel Fingerprint Recognition System with Direction Angles Difference

International Conference on Computational Intelligence and Multimedia Applications (ICCIMA 2007), 2007

... Maya V. Karki, Dr. S. Sethu Selvi Department of Electronics and Communication MS Ramaiah Inst... more

An Off Line Cursive Script Recognition System Using Fourier -Wavelet Features

International Conference on Computational Intelligence and Multimedia Applications (ICCIMA 2007), 2007

... Future work can be seen into incorporating neural network in the proposed algorithm. Acknowle... more

Andha Asthra-A navigation system for the visually impaired

Multisensor Fusion and …, 2008

... ANDHA ASTHRA# - A Navigation System for the Visually Impaired Dr. S. Sethu Selvi Umanath. R. ... more

An Offline Signature Recognition and Verification System Based on Neural Network

International Journal of Research in Engineering and Technology, 2014

Various techniques are already introduced for personal identification and verification based on d... more Various techniques are already introduced for personal identification and verification based on different types of biometrics which can be physiological or behavioral. Signatures lies in the category of behavioral biometric which can distort or changed with course of time. Signatures are considered to be most promising authentication method in all legal and financial documents. It is necessary to verify signers and their respective signatures. This paper presents an Offline Signature recognition and verification system(SRVS). In this system signature database of signature images is created, followed by image preprocessing, feature extraction, neural network design and training, and classification of signature as genuine or counterfeit.

Image Algebra and Image Fusion

Data Fusion Mathematics, 2015

Crowd Recognition System Based on Optical Flow Along with SVM classifier

International Journal of Electrical and Computer Engineering (IJECE), 2019

The manuscript discusses about abnormalities in a crowded scenario. To prevent the mishap at a pu... more The manuscript discusses about abnormalities in a crowded scenario. To prevent the mishap at a public place, there is no much mechanism which could prevent or alert the concerned authority about suspects in a crowd. Usually in a crowded scene, there are chances of some mishap like a terrorist attack or a crime. Our target is finding techniques to identify such activities and to possibly prevent them. If the crowd members exhibit abnormal behavior, we could identify and say that this particular person is a suspect and then the concerned authority would look into the matter. There are various methods to identify the abnormal behavior. The proposed approach is based on optical flow model. It has an ability to detect the sudden changes in motion of an individual among the crowd. First, the main region of motion is extracted by the help of motion heat map. Harris corner detector is used for extracting point of interest of extracted motion area. Based on the point of interest an optical f...

Geetha K.S, Pushpa M.K, Uttarakumari M & S.Sethu Selvi An Efficient Multiplierless Transform Algorithm for Video Coding

This paper presents an efficient algorithm to accelerate software video encoders/decoders by redu... more This paper presents an efficient algorithm to accelerate software video encoders/decoders by reducing the number of arithmetic operations for Discrete Cosine Transform (DCT). A multiplierless Ramanujan Ordered Number DCT (RDCT) is presented which computes the coefficients using shifts and addition operations only. The reduction in computational complexity has improved the performance of the video codec by almost 58 % compared with the commonly used integer DCT. The results show that significant computation reduction can be achieved with negligible average peak signal-to-noise ratio (PSNR) degradation. The average structural similarity index matrix (SSIM) also ensures that the degradation due to the approximation is minimal. Keywords: Ramanujan Ordered Number DCT, Multiplierless DCT, Video Coding. 1.

Variable Dimension Vector Quantization Based Image Watermarking

A watermarking methodbased on variable dimension vector quantization for hiding information in im... more A watermarking methodbased on variable dimension vector quantization for hiding information in imagesispresented in this paper. Watermark bits are embedded in the dimension information of the variable dimension reconstruction blocks of the cover or input image. Watermark extraction for oblivious watermarking is carried out by identifying the dimension of the reconstruction block. A variation of the scheme for cover escrow watermarking is also presentedtoincrease robustness. The simulation results prove the effectiveness of the proposed scheme and this scheme gives comparable capacity with the existing schemes.

Adaptive Square-Diamond Search (ASDS) Algorithm for Fast Block Matching Motion Estimation

This paper presents an efficient search algorithm for fast block matching motion estimation(ME) f... more This paper presents an efficient search algorithm for fast block matching motion estimation(ME) for video applications. The algorithm uses two patterns for initial search and refined local search. For initial search, the proposed algorithm uses a square pattern adaptively by selecting the step size based on Maximum Absolute Value of predicted motion vector. This search pattern is aimed to reduce the computational complexity of the ME block and find the least error. If the least error point is in the middle of the pattern, it shows that image is still and terminates the search. If the least error point is other than the middle point, then it becomes a new origin for subsequent refined local search with the pattern as small diamond. This is iteratively continued until the final motion vector is found. The efficacy of the proposed algorithm is verified by comparing with existing search algorithms in terms of PSNR, computations/search points and elapsed time for motion estimation. Simul...

Aircraft Parameter Estimation using Feedforward Neural Networks With Lyapunov Stability Analysis

Defence Science Journal, Nov 1, 2022

Aerodynamic parameter estimation is critical in the aviation sector, especially in design and dev... more Aerodynamic parameter estimation is critical in the aviation sector, especially in design and development programs of defense-military aircraft. In this paper, new results of the application of Artificial Neural Networks (ANN) to the field of aircraft parameter estimation are presented. The performances of Feedforward Neural Network (FFNN) with Backpropagation and FFNN with Backpropagation using Recursive Least Square (RLS) are investigated for aerodynamic parameter estimation. The methods are validated on flight data simulated using MATLAB implementations. The normalized Lyapunov energy functional has been used to derive the convergence conditions for both the ANN-based estimation algorithms. The estimation results are compared on the basis of performance metrics and computation time. The performance of FFNN-RLS has been observed to be approximately 10% better than FFNN-BPN. Simulation results from both algorithms have been found to be highly satisfactory and pave the way for further applications to real flight test data.

A Novel Deep Learning Algorithm for Covid Detection and Classification

2022 2nd International Conference on Intelligent Technologies (CONIT), Jun 24, 2022

Setnet: A Sparse Ensemble Network For Drone Localization And Zero Shot Drone Tracking In Real Time Surveillance Videos

2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)

Time Series based Models for Corona Data Analytics

2022 Fourth International Conference on Emerging Research in Electronics, Computer Science and Technology (ICERECT)

Machine Learning Based Classification of Welded Components

2022 4th International Conference on Circuits, Control, Communication and Computing (I4C)

A Deep Learning Approach to Classify Drones and Birds

2022 IEEE 2nd Mysore Sub Section International Conference (MysuruCon)

Deep Learning Approach to Nailfold Capillaroscopy Based Diabetes Mellitus Detection

International Journal of Online and Biomedical Engineering (iJOE)

Diabetes mellitus is a commonly occurring chronic metabolic disorder which has affected almost 40... more Diabetes mellitus is a commonly occurring chronic metabolic disorder which has affected almost 400 million people around the world. It can lead to vascular structure alterations and various renal, cardiovascular, and neurologicalcomplications claiming several lives. Since diabetes mellitus results is vascular structure changes, NailfoldCapillaroscopy(NFC) based approach can be employed for the detection of diabetes. NFC is an inexpensive, non-invasive method which involves acquisition of images of capillaries in the nail bed region using a USB digital microscope. Qualitative parameters of the capillaries such as tortuosity, hemorrhages, angiogenesis, elongated capillariesand quantitative parameters like length, width and mean capillary density are considered for diabetes detection. About 600 capillary images of healthy and diabetic subjects were collected and further data augmentation was performed to increase this to 1018 images dataset. This paper focuses on using NFC to obtain ca...

Effective Abnormal Behavior identification in a Crowd

2018 3rd International Conference on Circuits, Control, Communication and Computing (I4C), 2018

Effective analysis of crowd behavior is very challenging issue in video surveillance system. Most... more Effective analysis of crowd behavior is very challenging issue in video surveillance system. Most of the conventional approach is not up to the mark for real time behavior analysis. In this work, we have proposed an optical flow method along with KNCN classification approach to recognize abnormality in a video scene. Optical flow is a global motion information method which works efficiently. Optical flow vector provides direction and displacement of a point between two consecutive frames. This feature vector acts as input for KNN classifier. A surrounding neighbor (SN) concept is used for finding k-Nearest Centroid Neighbors(k-NCN) for each frame. We assign the class level based on the votes of k-NCN. The obtained result is promising and seems very efficient in terms of accuracy. We analyzed result and shown frame level comparison with ground truth. ROC analysis provides area under curve (AUC) for different data sequences and is compared with other state of the art techniques. Our proposed KNCN-OF method outperforms all other state of the art techniques with an accuracy of 99%.

Line Segmentation of Handwritten Kannada Text

Text line segmentation is an essential pre-processing stage for off-line handwriting recognition ... more Text line segmentation is an essential pre-processing stage for off-line handwriting recognition in many Optical Character Recognition (OCR) systems. It is an important step, as inaccurately segmented text lines will cause errors in the subsequent recognition stage. Text line segmentation of handwritten documents is still one of the most complicated problems in developing a reliable OCR. The nature of handwriting styles makes the process of text line segmentation very challenging. This paper provides a means of text line segmentation using Vertical Strip Projection based method. In this paper, the document is divided into vertical strips of appropriate width. Strip-wise horizontal histograms are then computed and the relationship of the peak valley points is used for line segmentation. Though the global horizontal projection method works efficiently for printed documents, it becomes highly unreliable when used in case of unconstrained handwritten text because of many factors like variable skew, touching and overlapping lines. The proposed method provides much improved and reliable results when compared to global horizontal profile method in case of skewed and variable sloping lines. Segmentation accuracy of 83% is achieved for handwritten Kannada text using the proposed method.

Multi-Algorithm Fusion for Fingerprint Recognition based on Texture Features

Establishing the identity of a person with highconfidence using biometric systems are gaining imp... more Establishing the identity of a person with highconfidence using biometric systems are gaining importance. It is achallenge to improve the recognition rate of an existing unimodalbiometric system. Fingerprint recognition is one of the most matureand proven technology because of its immutability and individuality.Recognition result of the system is based mainly on feature extractionmethod and type of matcher used. This paper proposes amulti-algorithm fusion algorithm for fingerprint recognition. The mainobjective of the proposed system is to improve performance usingtexture features. Feature extraction is based on ridge information andtextures. Orientation features, Curvelet transform features and DualTree-Complex Wavelet Transform (DT-CWT) features are extractedand using Euclidean distance match scores are evaluated. Texturefeatures need very less pre-processing compared to orientationfeatures. With this speed of the recognition system is improved.Weighted sum method is used in fusion of matchers. Performance ofindividual matchers in terms of False Acceptance Rate (FAR) andFalse Reject Rate (FRR) has been evaluated. For optimal threshold (η), percentage genuine recognition rate (%GAR) is calculated. Algorithm is tested on fingerprint database of 100 users and also with FVC2002-DB3 database. Maximum recognition rate of 95.2% is achieved by combining Curvelet and DT-CWT features

Blind Image Seperation Using Forward Difference Method (FDM)

Signal & Image Processing : An International Journal, 2011

In this paper, blind image separation is performed, exploiting the property of sparseness to repr... more In this paper, blind image separation is performed, exploiting the property of sparseness to represent images. A new sparse representation called forward difference method is proposed. It is known that most of the independent component analysis (ICA) basis functions, extracted from images are sparse and gives unreliable sparseness measure. In the proposed method, the image mixture is first transformed to sparse images. These images are divided into blocks and for each block the sparseness measure 0 norm is applied. The block having the most sparseness is considered to determine the separation matrix. The efficiency of the proposed method is compared with other sparse representation functions.

A Novel Fingerprint Recognition System with Direction Angles Difference

International Conference on Computational Intelligence and Multimedia Applications (ICCIMA 2007), 2007

... Maya V. Karki, Dr. S. Sethu Selvi Department of Electronics and Communication MS Ramaiah Inst... more

An Off Line Cursive Script Recognition System Using Fourier -Wavelet Features

International Conference on Computational Intelligence and Multimedia Applications (ICCIMA 2007), 2007

... Future work can be seen into incorporating neural network in the proposed algorithm. Acknowle... more

Andha Asthra-A navigation system for the visually impaired

Multisensor Fusion and …, 2008

... ANDHA ASTHRA# - A Navigation System for the Visually Impaired Dr. S. Sethu Selvi Umanath. R. ... more

An Offline Signature Recognition and Verification System Based on Neural Network

International Journal of Research in Engineering and Technology, 2014

Various techniques are already introduced for personal identification and verification based on d... more Various techniques are already introduced for personal identification and verification based on different types of biometrics which can be physiological or behavioral. Signatures lies in the category of behavioral biometric which can distort or changed with course of time. Signatures are considered to be most promising authentication method in all legal and financial documents. It is necessary to verify signers and their respective signatures. This paper presents an Offline Signature recognition and verification system(SRVS). In this system signature database of signature images is created, followed by image preprocessing, feature extraction, neural network design and training, and classification of signature as genuine or counterfeit.

Image Algebra and Image Fusion

Data Fusion Mathematics, 2015

Crowd Recognition System Based on Optical Flow Along with SVM classifier

International Journal of Electrical and Computer Engineering (IJECE), 2019

The manuscript discusses about abnormalities in a crowded scenario. To prevent the mishap at a pu... more The manuscript discusses about abnormalities in a crowded scenario. To prevent the mishap at a public place, there is no much mechanism which could prevent or alert the concerned authority about suspects in a crowd. Usually in a crowded scene, there are chances of some mishap like a terrorist attack or a crime. Our target is finding techniques to identify such activities and to possibly prevent them. If the crowd members exhibit abnormal behavior, we could identify and say that this particular person is a suspect and then the concerned authority would look into the matter. There are various methods to identify the abnormal behavior. The proposed approach is based on optical flow model. It has an ability to detect the sudden changes in motion of an individual among the crowd. First, the main region of motion is extracted by the help of motion heat map. Harris corner detector is used for extracting point of interest of extracted motion area. Based on the point of interest an optical f...

Geetha K.S, Pushpa M.K, Uttarakumari M & S.Sethu Selvi An Efficient Multiplierless Transform Algorithm for Video Coding

This paper presents an efficient algorithm to accelerate software video encoders/decoders by redu... more This paper presents an efficient algorithm to accelerate software video encoders/decoders by reducing the number of arithmetic operations for Discrete Cosine Transform (DCT). A multiplierless Ramanujan Ordered Number DCT (RDCT) is presented which computes the coefficients using shifts and addition operations only. The reduction in computational complexity has improved the performance of the video codec by almost 58 % compared with the commonly used integer DCT. The results show that significant computation reduction can be achieved with negligible average peak signal-to-noise ratio (PSNR) degradation. The average structural similarity index matrix (SSIM) also ensures that the degradation due to the approximation is minimal. Keywords: Ramanujan Ordered Number DCT, Multiplierless DCT, Video Coding. 1.

Variable Dimension Vector Quantization Based Image Watermarking

A watermarking methodbased on variable dimension vector quantization for hiding information in im... more A watermarking methodbased on variable dimension vector quantization for hiding information in imagesispresented in this paper. Watermark bits are embedded in the dimension information of the variable dimension reconstruction blocks of the cover or input image. Watermark extraction for oblivious watermarking is carried out by identifying the dimension of the reconstruction block. A variation of the scheme for cover escrow watermarking is also presentedtoincrease robustness. The simulation results prove the effectiveness of the proposed scheme and this scheme gives comparable capacity with the existing schemes.

Adaptive Square-Diamond Search (ASDS) Algorithm for Fast Block Matching Motion Estimation

This paper presents an efficient search algorithm for fast block matching motion estimation(ME) f... more This paper presents an efficient search algorithm for fast block matching motion estimation(ME) for video applications. The algorithm uses two patterns for initial search and refined local search. For initial search, the proposed algorithm uses a square pattern adaptively by selecting the step size based on Maximum Absolute Value of predicted motion vector. This search pattern is aimed to reduce the computational complexity of the ME block and find the least error. If the least error point is in the middle of the pattern, it shows that image is still and terminates the search. If the least error point is other than the middle point, then it becomes a new origin for subsequent refined local search with the pattern as small diamond. This is iteratively continued until the final motion vector is found. The efficacy of the proposed algorithm is verified by comparing with existing search algorithms in terms of PSNR, computations/search points and elapsed time for motion estimation. Simul...