Vijayan Asari | University of Dayton (original) (raw)

Papers by Vijayan Asari

Research paper thumbnail of Residues in succession recurrent U-Net for segmentation of retinal blood vessels

Research paper thumbnail of R2UNet for melt pond detection

Research paper thumbnail of Learning spatio-temporal attention for multi-object tracking and re-identification in wide area motion imagery

Research paper thumbnail of Histogram of Oriented Directional Features for Robust Face Recognition

International journal of monitoring and surveillance technologies research, Jul 1, 2016

Research paper thumbnail of GlacierNet2: A Hybrid Multi-Model Learning Architecture for Alpine Glacier Mapping

International journal of applied earth observation and geoinformation, Apr 6, 2022

Research paper thumbnail of Face recognition based on modular histogram of oriented directional features

This paper presents an illumination invariant face recognition system that uses local directional... more This paper presents an illumination invariant face recognition system that uses local directional pattern descriptor and modular histogram. The proposed Modular Histogram of Oriented Directional Features (MHODF) is an oriented local descriptor that is able to encode various patterns of face images under different lighting conditions. It employs the edge response values in different directions to encode each sub-image texture and produces multi-region histograms for each image. The edge responses are very important and play the main role for improving the face recognition accuracy. Therefore, we present the effectiveness of using different directional masks for detecting the edge responses on face recognition accuracy, such as Prewitt kernels, Kirsch masks, Sobel kernels, and Gaussian derivative masks. The performance evaluation of the proposed MHODF algorithm is conducted on several publicly available databases and observed promising recognition rates.

Research paper thumbnail of Automated point cloud completion for occlusion reduction in aerial lidar

Geospatial Informatics XII

Research paper thumbnail of Towards a deep-learning aided point cloud labeling suite

Pattern Recognition and Tracking XXXIII, Jun 13, 2022

Research paper thumbnail of A multimodal data registration strategy for 2D/3D information fusion

Multimodal Image Exploitation and Learning 2022, May 27, 2022

Research paper thumbnail of Deep neural machine for multimodal information fusion

Pattern Recognition and Tracking XXXIII, Jun 13, 2022

Research paper thumbnail of Residues in Succession U-Net for Fast and Efficient Segmentation

Pattern Recognition and Tracking XXXIII, Jun 13, 2022

Research paper thumbnail of Learning classical image registration features using a deep learning architecture

Pattern Recognition and Tracking XXXIII

Research paper thumbnail of Attention Focused Generative Network for Reducing Self-Occlusions in Aerial LiDAR

2021 IEEE Applied Imagery Pattern Recognition Workshop (AIPR)

This work explores the problem of re-ducing self-occlusion in aerial lidar. We introduce an entir... more This work explores the problem of re-ducing self-occlusion in aerial lidar. We introduce an entirely new dataset, DALES Viewpoints which uses a combination of synthetic and real-world data to represent aerial lidar scenes with different levels of occlusions. Our overall goal is to transform these occluded point clouds into a more visually complete representation of the same scene. We also propose a method called Channel Learned Downsampling (CLD). This downsampling method can be used as a drop-in replacement for any sampling method and uses a channel attention mechanism to select points based on their features instead of relying entirely on their spatial information. We show that this sampling method outperforms other sampling methods when used with a state-of-the-art point cloud completion network.

Research paper thumbnail of Long Short Working Memory (LSWM) Integration with Polynomial Connectivity for Object Tracking in Wide Area Motion Imagery

2018 15th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2018

High-value-target tracking in full-motion-video is a difficult surveillance task due to model dri... more High-value-target tracking in full-motion-video is a difficult surveillance task due to model drift while online training. We propose a tracker that adaptively fuses detections from multiple target models trained using three memory types to overcome model drift and other challenges. The short-term memory uses correlation and histogram features to detect the target from its recent appearance. The working memory uses a deep extreme learning network with polynomial connectivity that is trained online using a set of target appearances and background from the recent past. The long-term memory uses an offline trained polynomial CNN as a vehicle detector. Occlusion detection along with image registration and motion estimation stages aid in tracking the target through occlusions. A motion detection module reduces the effects of model drift and the scale change detector keeps the boundary accurate to the target. The proposed tracker is evaluated against state-of-the-art tracking methods on a vehicle dataset that includes challenging scenarios such as tree canopy occlusion, shadow, and erratic vehicle motion.

Research paper thumbnail of Multi-feature fusion based approach for robust face recognition

How to describe an image accurately with the most useful information is the key issue of any face... more How to describe an image accurately with the most useful information is the key issue of any face recognition task. Therefore, finding efficient and discriminative facial information that should be stable under different conditions of the image acquisition process is a huge challenge. Most existing approaches use only one type of features. In this paper, we argue that a robust face recognition technique requires several different kinds of information to be taken into account, suggesting the incorporation of several feature sets into a single fused one. Therefore, a new technique that combines the facial shape with the local structure and texture of the face image is proposed, namely multi-feature fusion (MFF). It is based on local boosted features (LBF) and Gabor wavelets techniques. Given an input image, the LBF histogram and Gabor features histogram are built separately. Then a final MFF feature descriptor is formed by concatenating these three histograms, which feeds to the suppo...

Research paper thumbnail of Multi-task deep learning architecture for semantic segmentation in EO imagery (Conference Presentation)

Pattern Recognition and Tracking XXX, 2019

Research paper thumbnail of Directional ringlet intensity feature transform (DRIFT) based object tracking in IR imagery

Current object tracking implementations utilize different feature extraction techniques to obtain... more Current object tracking implementations utilize different feature extraction techniques to obtain salient features to track objects of interest which change in different types of imaging modalities and environmental conditions.nChallenges in infrared imagery for object tracking include object deformation, occlusion, background variations, and smearing, which demands high performance algorithms. We propose the directional ringlet intensity feature transform to encompass significant levels of detail while being able to track low resolution targets. The algorithm utilizes a weighted circularly partitioned histogram distribution method which outperforms regular histogram distribution matching by localizing information and utilizing the rotation invariance of the circular rings. The image also utilizes directional edge information created by a Frei-Chen edge detector to improve the ability of the algorithm in different lighting conditions. We find the matching features using a weighted E...

Research paper thumbnail of Deep neural network based approach for robust aerial surveillance

Aerial object detection is one of the most important applications in computer vision. We propose ... more Aerial object detection is one of the most important applications in computer vision. We propose a deep learning strategy for detection and classification of objects on the pipeline right of ways by analyzing aerial images captured by flying aircrafts or drones. Due to the limitation of sufficient aerial datasets for accurately training the deep learning systems, it is necessary to create an efficient methodology for object data augmentation of the training dataset to achieve robust performance in various environmental conditions. Another limitation is the computing hardware that could be installed on the aircraft, especially when it is a drone. Hence a balance between the effectiveness and efficiency of object detector needs to be considered. We propose an efficient weighted IOU NMS (intersection over union non-maxima suppression) method to speed up the post-processing time that satisfies the onboard processing requirement. Weighted IOU NMS utilizes confidence scores of all propose...

Research paper thumbnail of Weighted Nonlinear Line Attractor for Complex Manifold Learning

Studies in Computational Intelligence, 2016

An artificial neural network is modeled by weighting between different neurons to form synaptic c... more An artificial neural network is modeled by weighting between different neurons to form synaptic connections. The nonlinear line attractor (NLA) models the weighting architecture by a polynomial weight set, which provides stronger connections between neurons. With the connections between neurons, we desired neuron weighting based on proximity using a Gaussian weighting strategy of the neurons that should reduce computational times significantly. Instead of using proximity to the neurons, it is found that utilizing the error found from estimating the output neurons to weight the connections between the neurons would provide the best results. The polynomial weights that are trained into the neural network will be reduced using a nonlinear dimensionality reduction which preserves the locality of the weights, since the weights are Gaussian weighted. A distance measure is then used to compare the test and training data. From testing the algorithm, it is observed that the proposed weighted NLA algorithm provides better recognition than both the GNLA algorithm and the original NLA algorithm.

Research paper thumbnail of Active Recall Networks for Multiperspectivity Learning through Shared Latent Space Optimization

Proceedings of the 11th International Joint Conference on Computational Intelligence, 2019

Given that there are numerous amounts of unlabeled data available for usage in training neural ne... more Given that there are numerous amounts of unlabeled data available for usage in training neural networks, it is desirable to implement a neural network architecture and training paradigm to maximize the ability of the latent space representation. Through multiple perspectives of the latent space using adversarial learning and autoencoding, data requirements can be reduced, which improves learning ability across domains. The entire goal of the proposed work is not to train exhaustively, but to train with multiperspectivity. We propose a new neural network architecture called Active Recall Network (ARN) for learning with less labels by optimizing the latent space. This neural network architecture learns latent space features of unlabeled data by using a fusion framework of an autoencoder and a generative adversarial network. Variations in the latent space representations will be captured and modeled by generation, discrimination, and reconstruction strategies in the network using both unlabeled and labeled data. Performance evaluations conducted on the proposed ARN architectures with two popular datasets demonstrated promising results in terms of generative capabilities and latent space effectiveness. Through the multiple perspectives that are embedded in ARN, we envision that this architecture will be incredibly versatile in every application that requires learning with less labels.

Research paper thumbnail of Residues in succession recurrent U-Net for segmentation of retinal blood vessels

Research paper thumbnail of R2UNet for melt pond detection

Research paper thumbnail of Learning spatio-temporal attention for multi-object tracking and re-identification in wide area motion imagery

Research paper thumbnail of Histogram of Oriented Directional Features for Robust Face Recognition

International journal of monitoring and surveillance technologies research, Jul 1, 2016

Research paper thumbnail of GlacierNet2: A Hybrid Multi-Model Learning Architecture for Alpine Glacier Mapping

International journal of applied earth observation and geoinformation, Apr 6, 2022

Research paper thumbnail of Face recognition based on modular histogram of oriented directional features

This paper presents an illumination invariant face recognition system that uses local directional... more This paper presents an illumination invariant face recognition system that uses local directional pattern descriptor and modular histogram. The proposed Modular Histogram of Oriented Directional Features (MHODF) is an oriented local descriptor that is able to encode various patterns of face images under different lighting conditions. It employs the edge response values in different directions to encode each sub-image texture and produces multi-region histograms for each image. The edge responses are very important and play the main role for improving the face recognition accuracy. Therefore, we present the effectiveness of using different directional masks for detecting the edge responses on face recognition accuracy, such as Prewitt kernels, Kirsch masks, Sobel kernels, and Gaussian derivative masks. The performance evaluation of the proposed MHODF algorithm is conducted on several publicly available databases and observed promising recognition rates.

Research paper thumbnail of Automated point cloud completion for occlusion reduction in aerial lidar

Geospatial Informatics XII

Research paper thumbnail of Towards a deep-learning aided point cloud labeling suite

Pattern Recognition and Tracking XXXIII, Jun 13, 2022

Research paper thumbnail of A multimodal data registration strategy for 2D/3D information fusion

Multimodal Image Exploitation and Learning 2022, May 27, 2022

Research paper thumbnail of Deep neural machine for multimodal information fusion

Pattern Recognition and Tracking XXXIII, Jun 13, 2022

Research paper thumbnail of Residues in Succession U-Net for Fast and Efficient Segmentation

Pattern Recognition and Tracking XXXIII, Jun 13, 2022

Research paper thumbnail of Learning classical image registration features using a deep learning architecture

Pattern Recognition and Tracking XXXIII

Research paper thumbnail of Attention Focused Generative Network for Reducing Self-Occlusions in Aerial LiDAR

2021 IEEE Applied Imagery Pattern Recognition Workshop (AIPR)

This work explores the problem of re-ducing self-occlusion in aerial lidar. We introduce an entir... more This work explores the problem of re-ducing self-occlusion in aerial lidar. We introduce an entirely new dataset, DALES Viewpoints which uses a combination of synthetic and real-world data to represent aerial lidar scenes with different levels of occlusions. Our overall goal is to transform these occluded point clouds into a more visually complete representation of the same scene. We also propose a method called Channel Learned Downsampling (CLD). This downsampling method can be used as a drop-in replacement for any sampling method and uses a channel attention mechanism to select points based on their features instead of relying entirely on their spatial information. We show that this sampling method outperforms other sampling methods when used with a state-of-the-art point cloud completion network.

Research paper thumbnail of Long Short Working Memory (LSWM) Integration with Polynomial Connectivity for Object Tracking in Wide Area Motion Imagery

2018 15th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2018

High-value-target tracking in full-motion-video is a difficult surveillance task due to model dri... more High-value-target tracking in full-motion-video is a difficult surveillance task due to model drift while online training. We propose a tracker that adaptively fuses detections from multiple target models trained using three memory types to overcome model drift and other challenges. The short-term memory uses correlation and histogram features to detect the target from its recent appearance. The working memory uses a deep extreme learning network with polynomial connectivity that is trained online using a set of target appearances and background from the recent past. The long-term memory uses an offline trained polynomial CNN as a vehicle detector. Occlusion detection along with image registration and motion estimation stages aid in tracking the target through occlusions. A motion detection module reduces the effects of model drift and the scale change detector keeps the boundary accurate to the target. The proposed tracker is evaluated against state-of-the-art tracking methods on a vehicle dataset that includes challenging scenarios such as tree canopy occlusion, shadow, and erratic vehicle motion.

Research paper thumbnail of Multi-feature fusion based approach for robust face recognition

How to describe an image accurately with the most useful information is the key issue of any face... more How to describe an image accurately with the most useful information is the key issue of any face recognition task. Therefore, finding efficient and discriminative facial information that should be stable under different conditions of the image acquisition process is a huge challenge. Most existing approaches use only one type of features. In this paper, we argue that a robust face recognition technique requires several different kinds of information to be taken into account, suggesting the incorporation of several feature sets into a single fused one. Therefore, a new technique that combines the facial shape with the local structure and texture of the face image is proposed, namely multi-feature fusion (MFF). It is based on local boosted features (LBF) and Gabor wavelets techniques. Given an input image, the LBF histogram and Gabor features histogram are built separately. Then a final MFF feature descriptor is formed by concatenating these three histograms, which feeds to the suppo...

Research paper thumbnail of Multi-task deep learning architecture for semantic segmentation in EO imagery (Conference Presentation)

Pattern Recognition and Tracking XXX, 2019

Research paper thumbnail of Directional ringlet intensity feature transform (DRIFT) based object tracking in IR imagery

Current object tracking implementations utilize different feature extraction techniques to obtain... more Current object tracking implementations utilize different feature extraction techniques to obtain salient features to track objects of interest which change in different types of imaging modalities and environmental conditions.nChallenges in infrared imagery for object tracking include object deformation, occlusion, background variations, and smearing, which demands high performance algorithms. We propose the directional ringlet intensity feature transform to encompass significant levels of detail while being able to track low resolution targets. The algorithm utilizes a weighted circularly partitioned histogram distribution method which outperforms regular histogram distribution matching by localizing information and utilizing the rotation invariance of the circular rings. The image also utilizes directional edge information created by a Frei-Chen edge detector to improve the ability of the algorithm in different lighting conditions. We find the matching features using a weighted E...

Research paper thumbnail of Deep neural network based approach for robust aerial surveillance

Aerial object detection is one of the most important applications in computer vision. We propose ... more Aerial object detection is one of the most important applications in computer vision. We propose a deep learning strategy for detection and classification of objects on the pipeline right of ways by analyzing aerial images captured by flying aircrafts or drones. Due to the limitation of sufficient aerial datasets for accurately training the deep learning systems, it is necessary to create an efficient methodology for object data augmentation of the training dataset to achieve robust performance in various environmental conditions. Another limitation is the computing hardware that could be installed on the aircraft, especially when it is a drone. Hence a balance between the effectiveness and efficiency of object detector needs to be considered. We propose an efficient weighted IOU NMS (intersection over union non-maxima suppression) method to speed up the post-processing time that satisfies the onboard processing requirement. Weighted IOU NMS utilizes confidence scores of all propose...

Research paper thumbnail of Weighted Nonlinear Line Attractor for Complex Manifold Learning

Studies in Computational Intelligence, 2016

An artificial neural network is modeled by weighting between different neurons to form synaptic c... more An artificial neural network is modeled by weighting between different neurons to form synaptic connections. The nonlinear line attractor (NLA) models the weighting architecture by a polynomial weight set, which provides stronger connections between neurons. With the connections between neurons, we desired neuron weighting based on proximity using a Gaussian weighting strategy of the neurons that should reduce computational times significantly. Instead of using proximity to the neurons, it is found that utilizing the error found from estimating the output neurons to weight the connections between the neurons would provide the best results. The polynomial weights that are trained into the neural network will be reduced using a nonlinear dimensionality reduction which preserves the locality of the weights, since the weights are Gaussian weighted. A distance measure is then used to compare the test and training data. From testing the algorithm, it is observed that the proposed weighted NLA algorithm provides better recognition than both the GNLA algorithm and the original NLA algorithm.

Research paper thumbnail of Active Recall Networks for Multiperspectivity Learning through Shared Latent Space Optimization

Proceedings of the 11th International Joint Conference on Computational Intelligence, 2019

Given that there are numerous amounts of unlabeled data available for usage in training neural ne... more Given that there are numerous amounts of unlabeled data available for usage in training neural networks, it is desirable to implement a neural network architecture and training paradigm to maximize the ability of the latent space representation. Through multiple perspectives of the latent space using adversarial learning and autoencoding, data requirements can be reduced, which improves learning ability across domains. The entire goal of the proposed work is not to train exhaustively, but to train with multiperspectivity. We propose a new neural network architecture called Active Recall Network (ARN) for learning with less labels by optimizing the latent space. This neural network architecture learns latent space features of unlabeled data by using a fusion framework of an autoencoder and a generative adversarial network. Variations in the latent space representations will be captured and modeled by generation, discrimination, and reconstruction strategies in the network using both unlabeled and labeled data. Performance evaluations conducted on the proposed ARN architectures with two popular datasets demonstrated promising results in terms of generative capabilities and latent space effectiveness. Through the multiple perspectives that are embedded in ARN, we envision that this architecture will be incredibly versatile in every application that requires learning with less labels.