Shankar Parmar - Academia.edu (original) (raw)

Papers by Shankar Parmar

2014 IEEE Geoscience and Remote Sensing Symposium, 2014

Successful outcomes of Sparse Representation-based Classifier (SRC) and Sparse Subspace Clusterin... more Successful outcomes of Sparse Representation-based Classifier (SRC) and Sparse Subspace Clustering (SSC) in many applications motivated us to combine these methods and propose a hierarchical classifier. The main idea behind the SRC and SSC algorithms is to represent a data using a sparse linear combination of elementary signals so that those elementary signals which are similar to the data contribute mainly in the representation. In this paper, the performance of a sparse representation based classifier is improved by pre-clustering of training samples using the SSC algorithm. A twostage SRC is then designed using the resulting clusters. A test data is classified by first determining the most similar cluster. The data label is subsequently found using the second stage classifier. The performance of the proposed method is evaluated considering cancer classification problem using the 14-Tumors microarray dataset. Due to low number of data samples per each class and high dimensionality of the data, this is a challenging problem. Curse of dimensionality, overfitting of the classifier to the training data and computational complexity are the possible related problems. Our experimental results show that the proposed method outperforms some other state of the art classifiers.

Healthcare analytics, Nov 1, 2023

The motivation of the proposed method is to solve typical problem for multiple object tracking li... more The motivation of the proposed method is to solve typical problem for multiple object tracking like partial background change, motion blur, object's occlusion, temporal movement, merging of object(s), etc. In this paper, we have proposed a color-based probability matching for real-time object(s) tracking. The proposed method is capable to detect moving object(s) and track the same object(s) which appear in the subsequent frame. Initially, object detection is carried by using two different approaches (i.e. Background Subtraction Modeling and Optical Flow Method) and pros and cons of both methods are discussed. Then, color-based probability is calculated for an individually detected object(s) in ensuing frames.

Mosaicing is the aligning of several frames into a single composition that represents part of a 3... more Mosaicing is the aligning of several frames into a single composition that represents part of a 3D scene. This is useful for many different applications, including virtual realty environments and movie special effects. Image mosaicing is commonly used to increase the visual field of view by pasting together many images or video frames. Mosaicing is a two step process. In the first step all the frames are registered individually. Image registration enables the geometric alignment of two images by means of homography. In this paper we have used optical fow method to find homography matrix between two frames. Second step is defined as the stitching of individually registered frames on a common canvas. Frames, which are considered in this paper, are views of a scene taken by a Pan, Tilt, and Zoom (PTZ) camera. Provided there is a sufficient overlap between frames, our's is a fully automatic and invariant to camera movement mosaicing technique.

Hyperspectral image contains hundreds of bands so it is spectrally overloaded and contains extent... more Hyperspectral image contains hundreds of bands so it is spectrally overloaded and contains extent information to differentiate spectrally unique material. Hyperspectral data generally used to identify the presence of material in scene. Almost all the hyperspectral cameras have spatial resolution limit (>5m per pixel) due to that each pixel can be a mixture of several materials. The process of unmixing is to unmixone of these mixed pixels. There are two models available to approximate mixing, (i) Linear Mixing Model (LMM) (ii) Nonlinear Mixing Model (NMM). Over a time, various approaches have been devised to address LMM and it's unmixing. In LMM, macrospectral mixtures are assumed. Nonlinear model comes under consideration due to microscopic mixing scale. In this paper, Generalized bilinear model is used which is nonlinear parametric model to get mixed data. Its accuracy depends on parametric form and parameter value chosen. It comes under convex optimization problem, so it can be solved using any optimization technique. Gradient descent algorithm (GDA) is employed to solve this optimization problem. Advantage of GDA over other unmixing techniques is that it transforms nonlinear model into linear one. To improve unmixing result, it is indeed advisable to consider spatial correlation among abundances. A novel approach has been introduced in this paper which considers 2nd order neighborhood correlation between abundances. Using our approach one can achieve better segmentation.

International Journal for Scientific Research and Development, Jul 1, 2014

Content Based Image Retrieval (CBIR) is emerging as important research in area of revolutionary i... more Content Based Image Retrieval (CBIR) is emerging as important research in area of revolutionary internet and digital technology. The focus of this paper is on efficient retrieval of similar images in a particular brain images using supporting vector machine (SVM). Instead of traditional low level features like color texture and shape which are uses in most of CBIR system Current approaches replace reshaped Image intensity as a feature to guide SVM system and applied to brain CT images. Original image is reshape to predefined size to limit the size of feature vector. Single matrix is generated as a database and class is assign to this matrix rows to train the SVM. This system helps radiologist to assist for evidence based practice or image based reasoning in his daily practice. Experimental results shows that the propose method is adequate and condescending to some other existing method

2022 IEEE 19th India Council International Conference (INDICON), Nov 24, 2022

Motion estimation is a key problem in digital video processing. In motion estimation, search patt... more Motion estimation is a key problem in digital video processing. In motion estimation, search patterns have a very important impact on searching speed and distortion performance. Hexagonal search achieves almost the same visual quality with full-search algorithm by taking fewer search points than diamond search. An adaptive threshold for early termination is introduced to avoid redundant calculation after the searching point is good enough. The simulation result shows that the proposed approach reduces the computational cost as compared to existing Hexagon Search Algorithm, with negligible loss in visual quality.

Discover Artificial Intelligence

Dyslexia is a learning disorder caused by difficulties in the brain’s processing of letters and w... more Dyslexia is a learning disorder caused by difficulties in the brain’s processing of letters and words. This study used EEG recordings to detect dyslexia at a young age. EEG recordings of 53 individuals, including 29 dyslexic and 24 normal individuals, were collected while they were engaged in two distinct mental activities known as the N-Back task and the Oddball task. Predictors were extracted using several methods and reduced using Principal Component Analysis (PCA). A relief-based strategy was applied to select predictors, and Support Vector Machine (SVM) classifier was used to achieve an average accuracy of 79.3% for dyslexia detection, which is better than the performance of its predecessors. The results indicate that EEG recordings and machine learning methods could be useful for identifying dyslexia in children.

2022 IEEE 19th India Council International Conference (INDICON)

Dyslexia is a neurological disorder affecting reading and writing abilities. Early intervention i... more Dyslexia is a neurological disorder affecting reading and writing abilities. Early intervention is important for affected individuals' social and academic development. The accuracy and objectivity limitations of traditional dyslexia detection systems based on behavioral symptoms and standard tests can pose challenges to the early detection of the condition. In response, an electroencephalogram (EEG) based detection method has been proposed to aid medical professionals in addressing these limitations. A comparison is made between the Wavelet Scattering Transform (WST) approach and three other approaches, namely Spectral Statistical Features (SSF), Connectivity Features with Autoencoders (CFA), and Hybrid Features (HF), using two datasets. These two datasets were chosen for various reasons, including the fact that they were collected during different tasks and from different countries. Another significant factor is that the age range of the participants was 7 and 12 years old, marking the beginning of their educational journey. This age range is ideal for detecting dyslexia in its primitive stages, making these datasets a perfect fit for this research. The performance evaluation of the approaches involved utilizing Support Vector Machine (SVM) classifiers with three non-linear kernels and k-fold cross-validation implementation. The findings suggest that the other three approaches could not achieve more than 80% accuracy, and their accuracy results were inconsistent with each dataset. In contrast, the WST approach achieved a high accuracy rate, with an average accuracy of 96.96% and 97.12% for dataset 1 and dataset 2, respectively, using the Radial Basis Function (RBF) kernel. The accuracy of WST features is further improved to 98.72% and 98.67% through the Majority Voting method. These findings demonstrate the effectiveness and generalization of the WST approach.

— Content Based Image Retrieval (CBIR) is emerging as important research in area of revolutionary... more — Content Based Image Retrieval (CBIR) is emerging as important research in area of revolutionary internet and digital technology. The focus of this paper is on efficient retrieval of similar images in a particular brain images using supporting vector machine (SVM). Instead of traditional low level features like color texture and shape which are uses in most of CBIR system Current approaches replace reshaped Image intensity as a feature to guide SVM system and applied to brain CT images. Original image is reshape to predefined size to limit the size of feature vector. Single matrix is generated as a database and class is assign to this matrix rows to train the SVM. This system helps radiologist to assist for evidence based practice or image based reasoning in his daily practice. Experimental results shows that the propose method is adequate and condescending to some other existing method

International Conference on Recent Advances and Innovations in Engineering (ICRAIE-2014), 2014

Motion estimation is a key problem in digital video processing. In motion estimation, search patt... more Motion estimation is a key problem in digital video processing. In motion estimation, search patterns have a very important impact on searching speed and distortion performance. Hexagonal search achieves almost the same visual quality with full-search algorithm by taking fewer search points than diamond search. An adaptive threshold for early termination is introduced to avoid redundant calculation after the searching point is good enough. The simulation result shows that the proposed approach reduces the computational cost as compared to existing Hexagon Search Algorithm, with negligible loss in visual quality.

2013 Nirma University International Conference on Engineering (NUiCONE), 2013

Colorization is an appealing area in the world of image processing. It has been used to increase ... more Colorization is an appealing area in the world of image processing. It has been used to increase the visual appeal of images such as old black and white photos, classic movies or scientific illustrations. Also, It has been very important in color image compression and video compression. Colorization, the task of coloring a grayscale image or video, involves assigning an intensity or luminance from the single dimension to a quantity that varies in three dimensions, such as red, green, and blue channels. Different colorization techniques are prevailing like hand coloring, semi automatic coloring and automatic coloring. Hand coloring and semi automatic coloring require intensive and perfect human interference. We are introducing a methodology for adding color to grayscale images in a way that is automatic and require little execution time. Towards this aim, a colorful image of the similar content as the grayscale image is taken, as an input source image by means of different image retrieval techniques. Then, the best matching source pixel is determined using luminance matching technique, for each pixel of the target grayscale image. Once a best matching source pixel is found, its chromaticity values are assigned to the target pixel while the original luminance value of the target pixel is retained.

Procedia Technology, 2014

Magnetic Resonance imaging (MRI) is a medical imaging procedure which uses strong magnetic fields... more Magnetic Resonance imaging (MRI) is a medical imaging procedure which uses strong magnetic fields and radio waves to produce cross sectional images of organs and internal structures of the body. Three dimensional (3D) models of CT is available and it has been practiced by almost all radiologists for pre-diagnosis. But in MRI still there is a scope for researcher to improvise a 3D model. Two dimensional images are taken from different viewpoints to reconstruct them in 3D, which is known as rendering process. In this paper, we have proposed a rendering concept for Medical (cardiac MRI) images based on iso values and number of marching cubes. Designer can place colors and textures over the 3D model to make it look realistic. This makes it easier for people to observe and visualize a substance in a better sense. The algorithm basically works on triangulation methods with various iso value and different combination of marching cube pairs. As a result of an application of marching cube concept, volumetric data (voxels) is generated. Voxels are then arranged and projected to visualize a 3D scene. Approximate processing time for various iso values are also compared in this paper.

International Conference on Recent Advances and Innovations in Engineering (ICRAIE-2014), 2014

In recent scenario, video surveillance systems are important to monitor security sensitive areas ... more In recent scenario, video surveillance systems are important to monitor security sensitive areas such as banks, crowded public places and borders. Hence, Object detection in such video surveillance is an important task. Research work to detect an object with steady background is carried out and upto some extent it is easy but with moving background it is still challenging. Object detection is complicated as the camera motion and object motion are mixed. Motion vectors, represents flow of a moving object, are obtained using Lucas - kanade optical flow algorithm for moving object detection with complex background. These flow vectors are quantized using a predefined threshold to decide whether a pixel is a part of an object or a background. We have used bilateral filter as a pre-processing step due to which accurate and fast detection of the object is achieved.

Differential methods of optical flow estimation are based on partial spatial and temporal derivat... more Differential methods of optical flow estimation are based on partial spatial and temporal derivatives of the image signal. In this paper, the comparison between background modeling technique and Lucas-Kanade optical flow has been done for object detection. Background subtraction methods need the background model from hundreds of images whereas the LucasKanade optical flow estimation method is a differential two frames algorithm, because it needs two frames in order to work. LucasKanade method is used which divides image into patches and computing a single optical flow on each of them. Keywords— Background Modeling, Motion Vector, Optical Flow, Object Detection

2014 IEEE Geoscience and Remote Sensing Symposium, 2014

Successful outcomes of Sparse Representation-based Classifier (SRC) and Sparse Subspace Clusterin... more Successful outcomes of Sparse Representation-based Classifier (SRC) and Sparse Subspace Clustering (SSC) in many applications motivated us to combine these methods and propose a hierarchical classifier. The main idea behind the SRC and SSC algorithms is to represent a data using a sparse linear combination of elementary signals so that those elementary signals which are similar to the data contribute mainly in the representation. In this paper, the performance of a sparse representation based classifier is improved by pre-clustering of training samples using the SSC algorithm. A twostage SRC is then designed using the resulting clusters. A test data is classified by first determining the most similar cluster. The data label is subsequently found using the second stage classifier. The performance of the proposed method is evaluated considering cancer classification problem using the 14-Tumors microarray dataset. Due to low number of data samples per each class and high dimensionality of the data, this is a challenging problem. Curse of dimensionality, overfitting of the classifier to the training data and computational complexity are the possible related problems. Our experimental results show that the proposed method outperforms some other state of the art classifiers.

Healthcare analytics, Nov 1, 2023

The motivation of the proposed method is to solve typical problem for multiple object tracking li... more The motivation of the proposed method is to solve typical problem for multiple object tracking like partial background change, motion blur, object's occlusion, temporal movement, merging of object(s), etc. In this paper, we have proposed a color-based probability matching for real-time object(s) tracking. The proposed method is capable to detect moving object(s) and track the same object(s) which appear in the subsequent frame. Initially, object detection is carried by using two different approaches (i.e. Background Subtraction Modeling and Optical Flow Method) and pros and cons of both methods are discussed. Then, color-based probability is calculated for an individually detected object(s) in ensuing frames.

Mosaicing is the aligning of several frames into a single composition that represents part of a 3... more Mosaicing is the aligning of several frames into a single composition that represents part of a 3D scene. This is useful for many different applications, including virtual realty environments and movie special effects. Image mosaicing is commonly used to increase the visual field of view by pasting together many images or video frames. Mosaicing is a two step process. In the first step all the frames are registered individually. Image registration enables the geometric alignment of two images by means of homography. In this paper we have used optical fow method to find homography matrix between two frames. Second step is defined as the stitching of individually registered frames on a common canvas. Frames, which are considered in this paper, are views of a scene taken by a Pan, Tilt, and Zoom (PTZ) camera. Provided there is a sufficient overlap between frames, our's is a fully automatic and invariant to camera movement mosaicing technique.

Hyperspectral image contains hundreds of bands so it is spectrally overloaded and contains extent... more Hyperspectral image contains hundreds of bands so it is spectrally overloaded and contains extent information to differentiate spectrally unique material. Hyperspectral data generally used to identify the presence of material in scene. Almost all the hyperspectral cameras have spatial resolution limit (>5m per pixel) due to that each pixel can be a mixture of several materials. The process of unmixing is to unmixone of these mixed pixels. There are two models available to approximate mixing, (i) Linear Mixing Model (LMM) (ii) Nonlinear Mixing Model (NMM). Over a time, various approaches have been devised to address LMM and it's unmixing. In LMM, macrospectral mixtures are assumed. Nonlinear model comes under consideration due to microscopic mixing scale. In this paper, Generalized bilinear model is used which is nonlinear parametric model to get mixed data. Its accuracy depends on parametric form and parameter value chosen. It comes under convex optimization problem, so it can be solved using any optimization technique. Gradient descent algorithm (GDA) is employed to solve this optimization problem. Advantage of GDA over other unmixing techniques is that it transforms nonlinear model into linear one. To improve unmixing result, it is indeed advisable to consider spatial correlation among abundances. A novel approach has been introduced in this paper which considers 2nd order neighborhood correlation between abundances. Using our approach one can achieve better segmentation.

International Journal for Scientific Research and Development, Jul 1, 2014

Content Based Image Retrieval (CBIR) is emerging as important research in area of revolutionary i... more Content Based Image Retrieval (CBIR) is emerging as important research in area of revolutionary internet and digital technology. The focus of this paper is on efficient retrieval of similar images in a particular brain images using supporting vector machine (SVM). Instead of traditional low level features like color texture and shape which are uses in most of CBIR system Current approaches replace reshaped Image intensity as a feature to guide SVM system and applied to brain CT images. Original image is reshape to predefined size to limit the size of feature vector. Single matrix is generated as a database and class is assign to this matrix rows to train the SVM. This system helps radiologist to assist for evidence based practice or image based reasoning in his daily practice. Experimental results shows that the propose method is adequate and condescending to some other existing method

2022 IEEE 19th India Council International Conference (INDICON), Nov 24, 2022

Motion estimation is a key problem in digital video processing. In motion estimation, search patt... more Motion estimation is a key problem in digital video processing. In motion estimation, search patterns have a very important impact on searching speed and distortion performance. Hexagonal search achieves almost the same visual quality with full-search algorithm by taking fewer search points than diamond search. An adaptive threshold for early termination is introduced to avoid redundant calculation after the searching point is good enough. The simulation result shows that the proposed approach reduces the computational cost as compared to existing Hexagon Search Algorithm, with negligible loss in visual quality.

Discover Artificial Intelligence

Dyslexia is a learning disorder caused by difficulties in the brain’s processing of letters and w... more Dyslexia is a learning disorder caused by difficulties in the brain’s processing of letters and words. This study used EEG recordings to detect dyslexia at a young age. EEG recordings of 53 individuals, including 29 dyslexic and 24 normal individuals, were collected while they were engaged in two distinct mental activities known as the N-Back task and the Oddball task. Predictors were extracted using several methods and reduced using Principal Component Analysis (PCA). A relief-based strategy was applied to select predictors, and Support Vector Machine (SVM) classifier was used to achieve an average accuracy of 79.3% for dyslexia detection, which is better than the performance of its predecessors. The results indicate that EEG recordings and machine learning methods could be useful for identifying dyslexia in children.

2022 IEEE 19th India Council International Conference (INDICON)

Dyslexia is a neurological disorder affecting reading and writing abilities. Early intervention i... more Dyslexia is a neurological disorder affecting reading and writing abilities. Early intervention is important for affected individuals' social and academic development. The accuracy and objectivity limitations of traditional dyslexia detection systems based on behavioral symptoms and standard tests can pose challenges to the early detection of the condition. In response, an electroencephalogram (EEG) based detection method has been proposed to aid medical professionals in addressing these limitations. A comparison is made between the Wavelet Scattering Transform (WST) approach and three other approaches, namely Spectral Statistical Features (SSF), Connectivity Features with Autoencoders (CFA), and Hybrid Features (HF), using two datasets. These two datasets were chosen for various reasons, including the fact that they were collected during different tasks and from different countries. Another significant factor is that the age range of the participants was 7 and 12 years old, marking the beginning of their educational journey. This age range is ideal for detecting dyslexia in its primitive stages, making these datasets a perfect fit for this research. The performance evaluation of the approaches involved utilizing Support Vector Machine (SVM) classifiers with three non-linear kernels and k-fold cross-validation implementation. The findings suggest that the other three approaches could not achieve more than 80% accuracy, and their accuracy results were inconsistent with each dataset. In contrast, the WST approach achieved a high accuracy rate, with an average accuracy of 96.96% and 97.12% for dataset 1 and dataset 2, respectively, using the Radial Basis Function (RBF) kernel. The accuracy of WST features is further improved to 98.72% and 98.67% through the Majority Voting method. These findings demonstrate the effectiveness and generalization of the WST approach.

— Content Based Image Retrieval (CBIR) is emerging as important research in area of revolutionary... more — Content Based Image Retrieval (CBIR) is emerging as important research in area of revolutionary internet and digital technology. The focus of this paper is on efficient retrieval of similar images in a particular brain images using supporting vector machine (SVM). Instead of traditional low level features like color texture and shape which are uses in most of CBIR system Current approaches replace reshaped Image intensity as a feature to guide SVM system and applied to brain CT images. Original image is reshape to predefined size to limit the size of feature vector. Single matrix is generated as a database and class is assign to this matrix rows to train the SVM. This system helps radiologist to assist for evidence based practice or image based reasoning in his daily practice. Experimental results shows that the propose method is adequate and condescending to some other existing method

International Conference on Recent Advances and Innovations in Engineering (ICRAIE-2014), 2014

Motion estimation is a key problem in digital video processing. In motion estimation, search patt... more Motion estimation is a key problem in digital video processing. In motion estimation, search patterns have a very important impact on searching speed and distortion performance. Hexagonal search achieves almost the same visual quality with full-search algorithm by taking fewer search points than diamond search. An adaptive threshold for early termination is introduced to avoid redundant calculation after the searching point is good enough. The simulation result shows that the proposed approach reduces the computational cost as compared to existing Hexagon Search Algorithm, with negligible loss in visual quality.

2013 Nirma University International Conference on Engineering (NUiCONE), 2013

Colorization is an appealing area in the world of image processing. It has been used to increase ... more Colorization is an appealing area in the world of image processing. It has been used to increase the visual appeal of images such as old black and white photos, classic movies or scientific illustrations. Also, It has been very important in color image compression and video compression. Colorization, the task of coloring a grayscale image or video, involves assigning an intensity or luminance from the single dimension to a quantity that varies in three dimensions, such as red, green, and blue channels. Different colorization techniques are prevailing like hand coloring, semi automatic coloring and automatic coloring. Hand coloring and semi automatic coloring require intensive and perfect human interference. We are introducing a methodology for adding color to grayscale images in a way that is automatic and require little execution time. Towards this aim, a colorful image of the similar content as the grayscale image is taken, as an input source image by means of different image retrieval techniques. Then, the best matching source pixel is determined using luminance matching technique, for each pixel of the target grayscale image. Once a best matching source pixel is found, its chromaticity values are assigned to the target pixel while the original luminance value of the target pixel is retained.

Procedia Technology, 2014

Magnetic Resonance imaging (MRI) is a medical imaging procedure which uses strong magnetic fields... more Magnetic Resonance imaging (MRI) is a medical imaging procedure which uses strong magnetic fields and radio waves to produce cross sectional images of organs and internal structures of the body. Three dimensional (3D) models of CT is available and it has been practiced by almost all radiologists for pre-diagnosis. But in MRI still there is a scope for researcher to improvise a 3D model. Two dimensional images are taken from different viewpoints to reconstruct them in 3D, which is known as rendering process. In this paper, we have proposed a rendering concept for Medical (cardiac MRI) images based on iso values and number of marching cubes. Designer can place colors and textures over the 3D model to make it look realistic. This makes it easier for people to observe and visualize a substance in a better sense. The algorithm basically works on triangulation methods with various iso value and different combination of marching cube pairs. As a result of an application of marching cube concept, volumetric data (voxels) is generated. Voxels are then arranged and projected to visualize a 3D scene. Approximate processing time for various iso values are also compared in this paper.

International Conference on Recent Advances and Innovations in Engineering (ICRAIE-2014), 2014

In recent scenario, video surveillance systems are important to monitor security sensitive areas ... more In recent scenario, video surveillance systems are important to monitor security sensitive areas such as banks, crowded public places and borders. Hence, Object detection in such video surveillance is an important task. Research work to detect an object with steady background is carried out and upto some extent it is easy but with moving background it is still challenging. Object detection is complicated as the camera motion and object motion are mixed. Motion vectors, represents flow of a moving object, are obtained using Lucas - kanade optical flow algorithm for moving object detection with complex background. These flow vectors are quantized using a predefined threshold to decide whether a pixel is a part of an object or a background. We have used bilateral filter as a pre-processing step due to which accurate and fast detection of the object is achieved.

Differential methods of optical flow estimation are based on partial spatial and temporal derivat... more Differential methods of optical flow estimation are based on partial spatial and temporal derivatives of the image signal. In this paper, the comparison between background modeling technique and Lucas-Kanade optical flow has been done for object detection. Background subtraction methods need the background model from hundreds of images whereas the LucasKanade optical flow estimation method is a differential two frames algorithm, because it needs two frames in order to work. LucasKanade method is used which divides image into patches and computing a single optical flow on each of them. Keywords— Background Modeling, Motion Vector, Optical Flow, Object Detection