nitigya sambyal - Academia.edu (original) (raw)

Uploads

Papers by nitigya sambyal

Research paper thumbnail of A Survey of Machine Learning- and Deep Learning-Based Techniques for Diabetic Retinopathy Screening

CRC Press eBooks, Dec 20, 2023

Research paper thumbnail of Modified residual networks for severity stage classification of diabetic retinopathy

Research paper thumbnail of Connected Component based English Character Set Segmentation

Character segmentation is a prerequisite phase in document analysis tasks. It helps in extracting... more Character segmentation is a prerequisite phase in document analysis tasks. It helps in extracting characters which aid in indexing and understanding. The research paper proposes a character segmentation technique based on connected component method. The algorithm is tested along a set of jpeg, png and bmp images over English character set and uses a cluster size of five for identifying the separate characters. It has been observed that the proposed connected component based character segmentation approach gives on average 95.44% accuracy for the various test cases under study. Keywords— Connected component; Character segmentation,; Normalization ;Thresholding

Research paper thumbnail of Automatic text extraction and character segmentation using maximally stable extremal regions

ArXiv, 2016

Text detection and segmentation is an important prerequisite for many content based image analysi... more Text detection and segmentation is an important prerequisite for many content based image analysis tasks. The paper proposes a novel text extraction and character segmentation algorithm using Maximally Stable Extremal Regions as basic letter candidates. These regions are then subjected to thresholding and thereafter various connected components are determined to identify separate characters. The algorithm is tested along a set of various JPEG, PNG and BMP images over four different character sets; English, Russian, Hindi and Urdu. The algorithm gives good results for English and Russian character set; however character segmentation in Urdu and Hindi language is not much accurate. The algorithm is simple, efficient, involves no overhead as required in training and gives good results for even low quality images. The paper also proposes various challenges in text extraction and segmentation for multilingual inputs.

Research paper thumbnail of Affective Computing: Challenges and Prospect

Affective computing simulates empathy through machine that could recognize, interpret, process an... more Affective computing simulates empathy through machine that could recognize, interpret, process and respond to human affects. With the use of sensors and computational devices, it proposes to exhibit either innate emotional capabilities or that is capable of convincingly simulating emotions. The paper focuses on varied challenges and future scope of affective computing. The technologies for affective computing are varied but expression if not natural may not yield 100% accurate results. The systems may lack rotational movement freedom and also ignores situational factors in emotional understanding. Keywords—Human computer interaction; emotions.

Research paper thumbnail of Aggregated residual transformation network for multistage classification in diabetic retinopathy

International Journal of Imaging Systems and Technology, 2020

Diabetic Retinopathy is a retinal abnormality which is characterized by progressive damage to the... more Diabetic Retinopathy is a retinal abnormality which is characterized by progressive damage to the retina, eventually leading to irreversible blindness. In this paper, we propose an aggregated residual transformation-based model for automatic multistage classification of diabetic retinopathy. The proposed model obtains 99.68% overall classification accuracy, 99.68% sensitivity, 99.89% specificity and 99.68% precision without overfitting on the MESSIDOR dataset. Further, the model obtains an accuracy of 99.89% for stage 0, 99.89% for stage 1, 99.68% for stage 2 and 99.89% for stage 3 of diabetic retinopathy. In comparison to residual network, the model shows an overall accuracy gain of 0.52%. The model also ensures an overall improvement of more than 6% in accuracy, 1.2% in sensitivity and 2.43 % in specificity when compared to best results reported in the literature. The proposed work outperforms the existing methods and achieves state-of-the-art results for the multistage classification of diabetic retinopathy.

Research paper thumbnail of Modified U-Net architecture for semantic segmentation of diabetic retinopathy images

Biocybernetics and Biomedical Engineering, 2020

Segmentation of lesions from fundus images is an essential prerequisite for accurate severity ass... more Segmentation of lesions from fundus images is an essential prerequisite for accurate severity assessment of diabetic retinopathy. Due to variation in morphologies, number and size of lesions, the manual grading process becomes extremely challenging and timeconsuming. This necessitates the need of an automatic segmentation system that can precisely define the region of interest boundaries and assist ophthalmologists in speedy diagnosis along with diabetic retinopathy severity grading. The paper presents a modified U-Net architecture based on residual network and employs periodic shuffling with sub-pixel convolution initialized to convolution nearest neighbour resize. The proposed architecture has been trained and validated for microaneurysm and hard exudate segmentation on two publicly available datasets namely IDRiD and e-ophtha. For IDRiD dataset, the network obtains 99.88% accuracy, 99.85% sensitivity, 99.95% specificity and dice score of 0.9998 for both microaneurysm and exudate segmentation. Further, when trained on e-ophtha and validated on IDRiD dataset, the network shows 99.98% accuracy, 99.88% sensitivity, 99.89% specificity and dice score of 0.9998 for microaneurysm segmentation. For exudates segmentation, the model obtains 99.98% accuracy, 99.88% sensitivity, 99.89% specificity and dice score of 0.9999 when trained on e-ophtha and validated on IDRiD dataset. In comparison to existing literature, the proposed model provides state-of-the-art results for retinal lesion segmentation.

Research paper thumbnail of Big Data Analytics

Advances in Computer and Electrical Engineering, 2019

Research paper thumbnail of Feature based Text Extraction System using Connected Component Method

International Journal of Synthetic Emotions, 2016

Text detection and segmentation system serves as important method for document analysis as it hel... more Text detection and segmentation system serves as important method for document analysis as it helps in many content based image analysis tasks. This research paper proposes a connected component technique for text extraction and character segmentation using maximally stable extremal regions (MSERs) for text line formation followed by connected components to determined separate characters. The system uses a cluster size of five which is selected by experimental evaluation for identifying characters. Sobel edge detector is used as it reduces the execution time but at the same time maintains quality of the results. The algorithm is tested along a set of JPEG, PNG and BMP images over varying features like font size, style, colour, background colour and text variation. Further the CPU time in execution of the algorithm with three different edge detectors namely prewitt, sobel and canny is observed. Text identification using MSER gave very good results whereas character segmentation gave ...

Research paper thumbnail of Restoration of artwork using deep neural networks

Evolving Systems, 2019

Paintings and other similar work of art represent an important part of our heritage and contempor... more Paintings and other similar work of art represent an important part of our heritage and contemporary culture. However, due to the nature of materials used in these works, they are prone to damage and degradation over a period of time. Some of damages to these works may include torn canvases, smudges, exposure to elements etc. This necessitates the need for restoration of artworks. The restoration process is very time consuming and is a delicate task making it prone to human error. The virtual restoration of digitized artworks can be very helpful in this process. In this paper, we have proposed a method based on deep neural networks for virtual restoration of the digitized artworks. The paper presents a hybrid model which employs automatic mask generation based on Mask R-CNN and image inpainting using U-Net architecture with partial convolutions and automatic mask update. The proposed approach is evaluated qualitatively as well as quantitatively. The qualitative evaluation of the approach is done by engaging three domain art experts. On the other hand, quantitative validation of the proposed method is done using dataset of images having artificially created irregular holes by employing mean square error (MSE) and structural similarity index (SSIM) metrics. The results obtained show that the proposed approach is quite effective in virtual restoration of the digitized artworks.

Research paper thumbnail of A Discriminative Learning-Based Deep Learning Approach for Diabetic Retinopathy Classification

Lecture notes in electrical engineering, 2022

Research paper thumbnail of A Review of Statistical and Machine Learning Techniques for Microvascular Complications in Type 2 Diabetes

Current diabetes reviews, 2020

Background and Introduction: Diabetes mellitus is a metabolic disorder that has emerged as a seri... more Background and Introduction: Diabetes mellitus is a metabolic disorder that has emerged as a serious public health issue worldwide. According to the World Health Organization (WHO), without interventions, the number of diabetic incidences is expected to be at least 629 million by 2045. Uncontrolled diabetes gradually leads to progressive damage to eyes, heart, kidneys, blood vessels and nerves. METHOD The paper presents a critical review of existing statistical and Artificial Intelligence (AI) based machine learning techniques with respect to DM complications namely retinopathy, neuropathy and nephropathy. The statistical and machine learning analytic techniques are used to structure the subsequent content review. RESULT It has been inferred that statistical analysis can help only in inferential and descriptive analysis whereas, AI based machine learning models can even provide actionable prediction models for faster and accurate diagnose of complications associated with DM. CONCLUS...

Research paper thumbnail of Breast cancer detection from histopathology images using modified residual neural networks

Biocybernetics and Biomedical Engineering

Research paper thumbnail of Microvascular Complications in Type-2 Diabetes: A Review of Statistical Techniques and Machine Learning Models

Wireless Personal Communications

Research paper thumbnail of Power Theft Detection Using Deep Neural Networks

Electric Power Components and Systems

Research paper thumbnail of Detection and localization of potholes in thermal images using deep neural networks

Multimedia Tools and Applications

Research paper thumbnail of Automatic text extraction and character segmentation using maximally stable extremal regions

Text detection and segmentation is an important prerequisite for many content based image analysi... more Text detection and segmentation is an important prerequisite for many content based image analysis tasks. The paper proposes a novel text extraction and character segmentation algorithm using Maximally Stable Extremal Regions as basic letter candidates. These regions are then subjected to thresholding and thereafter various connected components are determined to identify separate characters. The algorithm is tested along a set of various JPEG, PNG and BMP images over four different character sets; English, Russian, Hindi and Urdu. The algorithm gives good results for English and Russian character set; however character segmentation in Urdu and Hindi language is not much accurate. The algorithm is simple, efficient, involves no overhead as required in training and gives good results for even low quality images. The paper also proposes various challenges in text extraction and segmentation for multilingual inputs.

Research paper thumbnail of A Survey of Machine Learning- and Deep Learning-Based Techniques for Diabetic Retinopathy Screening

CRC Press eBooks, Dec 20, 2023

Research paper thumbnail of Modified residual networks for severity stage classification of diabetic retinopathy

Research paper thumbnail of Connected Component based English Character Set Segmentation

Character segmentation is a prerequisite phase in document analysis tasks. It helps in extracting... more Character segmentation is a prerequisite phase in document analysis tasks. It helps in extracting characters which aid in indexing and understanding. The research paper proposes a character segmentation technique based on connected component method. The algorithm is tested along a set of jpeg, png and bmp images over English character set and uses a cluster size of five for identifying the separate characters. It has been observed that the proposed connected component based character segmentation approach gives on average 95.44% accuracy for the various test cases under study. Keywords— Connected component; Character segmentation,; Normalization ;Thresholding

Research paper thumbnail of Automatic text extraction and character segmentation using maximally stable extremal regions

ArXiv, 2016

Text detection and segmentation is an important prerequisite for many content based image analysi... more Text detection and segmentation is an important prerequisite for many content based image analysis tasks. The paper proposes a novel text extraction and character segmentation algorithm using Maximally Stable Extremal Regions as basic letter candidates. These regions are then subjected to thresholding and thereafter various connected components are determined to identify separate characters. The algorithm is tested along a set of various JPEG, PNG and BMP images over four different character sets; English, Russian, Hindi and Urdu. The algorithm gives good results for English and Russian character set; however character segmentation in Urdu and Hindi language is not much accurate. The algorithm is simple, efficient, involves no overhead as required in training and gives good results for even low quality images. The paper also proposes various challenges in text extraction and segmentation for multilingual inputs.

Research paper thumbnail of Affective Computing: Challenges and Prospect

Affective computing simulates empathy through machine that could recognize, interpret, process an... more Affective computing simulates empathy through machine that could recognize, interpret, process and respond to human affects. With the use of sensors and computational devices, it proposes to exhibit either innate emotional capabilities or that is capable of convincingly simulating emotions. The paper focuses on varied challenges and future scope of affective computing. The technologies for affective computing are varied but expression if not natural may not yield 100% accurate results. The systems may lack rotational movement freedom and also ignores situational factors in emotional understanding. Keywords—Human computer interaction; emotions.

Research paper thumbnail of Aggregated residual transformation network for multistage classification in diabetic retinopathy

International Journal of Imaging Systems and Technology, 2020

Diabetic Retinopathy is a retinal abnormality which is characterized by progressive damage to the... more Diabetic Retinopathy is a retinal abnormality which is characterized by progressive damage to the retina, eventually leading to irreversible blindness. In this paper, we propose an aggregated residual transformation-based model for automatic multistage classification of diabetic retinopathy. The proposed model obtains 99.68% overall classification accuracy, 99.68% sensitivity, 99.89% specificity and 99.68% precision without overfitting on the MESSIDOR dataset. Further, the model obtains an accuracy of 99.89% for stage 0, 99.89% for stage 1, 99.68% for stage 2 and 99.89% for stage 3 of diabetic retinopathy. In comparison to residual network, the model shows an overall accuracy gain of 0.52%. The model also ensures an overall improvement of more than 6% in accuracy, 1.2% in sensitivity and 2.43 % in specificity when compared to best results reported in the literature. The proposed work outperforms the existing methods and achieves state-of-the-art results for the multistage classification of diabetic retinopathy.

Research paper thumbnail of Modified U-Net architecture for semantic segmentation of diabetic retinopathy images

Biocybernetics and Biomedical Engineering, 2020

Segmentation of lesions from fundus images is an essential prerequisite for accurate severity ass... more Segmentation of lesions from fundus images is an essential prerequisite for accurate severity assessment of diabetic retinopathy. Due to variation in morphologies, number and size of lesions, the manual grading process becomes extremely challenging and timeconsuming. This necessitates the need of an automatic segmentation system that can precisely define the region of interest boundaries and assist ophthalmologists in speedy diagnosis along with diabetic retinopathy severity grading. The paper presents a modified U-Net architecture based on residual network and employs periodic shuffling with sub-pixel convolution initialized to convolution nearest neighbour resize. The proposed architecture has been trained and validated for microaneurysm and hard exudate segmentation on two publicly available datasets namely IDRiD and e-ophtha. For IDRiD dataset, the network obtains 99.88% accuracy, 99.85% sensitivity, 99.95% specificity and dice score of 0.9998 for both microaneurysm and exudate segmentation. Further, when trained on e-ophtha and validated on IDRiD dataset, the network shows 99.98% accuracy, 99.88% sensitivity, 99.89% specificity and dice score of 0.9998 for microaneurysm segmentation. For exudates segmentation, the model obtains 99.98% accuracy, 99.88% sensitivity, 99.89% specificity and dice score of 0.9999 when trained on e-ophtha and validated on IDRiD dataset. In comparison to existing literature, the proposed model provides state-of-the-art results for retinal lesion segmentation.

Research paper thumbnail of Big Data Analytics

Advances in Computer and Electrical Engineering, 2019

Research paper thumbnail of Feature based Text Extraction System using Connected Component Method

International Journal of Synthetic Emotions, 2016

Text detection and segmentation system serves as important method for document analysis as it hel... more Text detection and segmentation system serves as important method for document analysis as it helps in many content based image analysis tasks. This research paper proposes a connected component technique for text extraction and character segmentation using maximally stable extremal regions (MSERs) for text line formation followed by connected components to determined separate characters. The system uses a cluster size of five which is selected by experimental evaluation for identifying characters. Sobel edge detector is used as it reduces the execution time but at the same time maintains quality of the results. The algorithm is tested along a set of JPEG, PNG and BMP images over varying features like font size, style, colour, background colour and text variation. Further the CPU time in execution of the algorithm with three different edge detectors namely prewitt, sobel and canny is observed. Text identification using MSER gave very good results whereas character segmentation gave ...

Research paper thumbnail of Restoration of artwork using deep neural networks

Evolving Systems, 2019

Paintings and other similar work of art represent an important part of our heritage and contempor... more Paintings and other similar work of art represent an important part of our heritage and contemporary culture. However, due to the nature of materials used in these works, they are prone to damage and degradation over a period of time. Some of damages to these works may include torn canvases, smudges, exposure to elements etc. This necessitates the need for restoration of artworks. The restoration process is very time consuming and is a delicate task making it prone to human error. The virtual restoration of digitized artworks can be very helpful in this process. In this paper, we have proposed a method based on deep neural networks for virtual restoration of the digitized artworks. The paper presents a hybrid model which employs automatic mask generation based on Mask R-CNN and image inpainting using U-Net architecture with partial convolutions and automatic mask update. The proposed approach is evaluated qualitatively as well as quantitatively. The qualitative evaluation of the approach is done by engaging three domain art experts. On the other hand, quantitative validation of the proposed method is done using dataset of images having artificially created irregular holes by employing mean square error (MSE) and structural similarity index (SSIM) metrics. The results obtained show that the proposed approach is quite effective in virtual restoration of the digitized artworks.

Research paper thumbnail of A Discriminative Learning-Based Deep Learning Approach for Diabetic Retinopathy Classification

Lecture notes in electrical engineering, 2022

Research paper thumbnail of A Review of Statistical and Machine Learning Techniques for Microvascular Complications in Type 2 Diabetes

Current diabetes reviews, 2020

Background and Introduction: Diabetes mellitus is a metabolic disorder that has emerged as a seri... more Background and Introduction: Diabetes mellitus is a metabolic disorder that has emerged as a serious public health issue worldwide. According to the World Health Organization (WHO), without interventions, the number of diabetic incidences is expected to be at least 629 million by 2045. Uncontrolled diabetes gradually leads to progressive damage to eyes, heart, kidneys, blood vessels and nerves. METHOD The paper presents a critical review of existing statistical and Artificial Intelligence (AI) based machine learning techniques with respect to DM complications namely retinopathy, neuropathy and nephropathy. The statistical and machine learning analytic techniques are used to structure the subsequent content review. RESULT It has been inferred that statistical analysis can help only in inferential and descriptive analysis whereas, AI based machine learning models can even provide actionable prediction models for faster and accurate diagnose of complications associated with DM. CONCLUS...

Research paper thumbnail of Breast cancer detection from histopathology images using modified residual neural networks

Biocybernetics and Biomedical Engineering

Research paper thumbnail of Microvascular Complications in Type-2 Diabetes: A Review of Statistical Techniques and Machine Learning Models

Wireless Personal Communications

Research paper thumbnail of Power Theft Detection Using Deep Neural Networks

Electric Power Components and Systems

Research paper thumbnail of Detection and localization of potholes in thermal images using deep neural networks

Multimedia Tools and Applications

Research paper thumbnail of Automatic text extraction and character segmentation using maximally stable extremal regions

Text detection and segmentation is an important prerequisite for many content based image analysi... more Text detection and segmentation is an important prerequisite for many content based image analysis tasks. The paper proposes a novel text extraction and character segmentation algorithm using Maximally Stable Extremal Regions as basic letter candidates. These regions are then subjected to thresholding and thereafter various connected components are determined to identify separate characters. The algorithm is tested along a set of various JPEG, PNG and BMP images over four different character sets; English, Russian, Hindi and Urdu. The algorithm gives good results for English and Russian character set; however character segmentation in Urdu and Hindi language is not much accurate. The algorithm is simple, efficient, involves no overhead as required in training and gives good results for even low quality images. The paper also proposes various challenges in text extraction and segmentation for multilingual inputs.