Papers by BIBEK R A N J A N GHOSH
Journal of Mathematical Sciences & Computational Mathematics, 2022
A brain tumor is an abnormal tissue growth in the brain in which cells grow and multiply uncontro... more A brain tumor is an abnormal tissue growth in the brain in which cells grow and multiply uncontrollably. The main aim of our task is to identify brain tumor from a particular MRI image and to extract the tumor region. The proposed method uses various digital image processing techniques such as noise removal, Otsu's thresholding, morphological operations etc. This paper focuses on the use of marker-based watershed transform for segmentation, to get near about accurate results.
Automated Face Recognition System by PCA and DWT Feature Engineering with Optimized Machine Learning Methods, 2022
The Principal Component Analysis (PCA) is a technique that reduces data dimension by synthesizing... more The Principal Component Analysis (PCA) is a technique that reduces data dimension by synthesizing important representative features. Discrete Wavelet Transform (DWT) is a strong signal processing tool that can analyze the image in multi-resolutions and can discover significant features like edges. Properly trained supervised machine learning models can recognize input face pretty correctly. In this work face features extracted using PCA and DWT are assembled together to train various Machine Learning Models. Each model is studied using four benchmark face data sets namely YALE, JAFEE, GEORGIA TECH and ORL. The models are further optimized and their performances compared with existing state of the art methods. For YALE the proposed method achieved 94.54% accuracy using Logistic regression, for JAFEE 99.5% accuracy in Logistic regression, for GEORGIA TECH and ORL the method also out performs the existing techniques by achieving 84.6% and 98.5% accuracies.
2022 International Conference on Inventive Computation Technologies (ICICT)
Digital watermarking has immense importance in protection of copyright as well as authentication.... more Digital watermarking has immense importance in protection of copyright as well as authentication. In this work, a watermarking process utilising S transform is proposed with Haar wavelet that generate only integer coefficients. Initially, the forward transform of cover image generates four coefficient bands-HH, HL, LH and LL. The HH sub-band location map is randomized and watermark image is embedded using least significant bits (LSB) substitution of the coefficient. Additionally 256 bit SHA hash code of the watermark image is generated for authentication and embedded in the HL sub-band in a similar fashion. During extraction, watermark is extracted and SHA256 hash code is computed again. This new hash code is compared with extracted hash code for authenticity. The average PSNR (Peak Signal to Noise Ratio) observed for SIPI dataset is 65 dB and 59 dB with 8192 bit and 32768 bit watermark respectively with embedding rate of 2 bits per pixel. The results of the proposed merthod are checked against existing methods and we found it to be superior in many cases.
A Deep Learning Based Image Steganalysis Using Gray Level Co-Occurrence Matrix
2022 Second International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies (ICAECT)
International Journal of Computer Sciences and Engineering, 2022
International Journal of Computer Sciences and Engineering, 2022
Electronic patient record (EPR) storage and transmission is an integral part of modern healthcare... more Electronic patient record (EPR) storage and transmission is an integral part of modern healthcare system.
Steganography comes into play when the medical image becomes a potential storage of patient information. Generally
steganographic techniques give higher importance to the perfect recovery of secret data than the actual reconstruction of the
original image. But in medical system the actual reconstruction of the original medical image is importance as this is the
primary source of disease understanding. Hence in this context reversible steganography plays an important role in
telemedicine field in storing, maintenance and communication of electronic patient record within the image itself, so that
where a doctor located at a geographically remote location can diagnose a patient on the basis of a medical image generated
elsewhere with confidence. In this paper we have studied some elegant techniques for reversible medical image
steganographic techniques and compare their performances.
Covid19 is the menace of this century. World Health Organization (WHO) declared it pandemic in Fe... more Covid19 is the menace of this century. World Health Organization (WHO) declared it pandemic in February, 2020. This RNA virus has catastrophic impact of the entire human civilization since it was initially reported to have been erupted from Wuhan, a city in Hubei province of China in late December 2019. In the first wave millions of people died in many countries. Even the developed countries like USA, France, Italy, United Kingdom etc. were in shock and could not prevent loss of human lives with their well-established medical infrastructure. Strict lockdown, quarantines were imposed. The hospitals were outnumbered by the severely ill patients who needed ventilation support. Many died without treatment, dead bodies were on the streets and mass graves became a practice. Developing and underdeveloped countries faced even more disastrous situations. Since then the virus is mutating and giving new challenges to human society in developing a cure. Until now RTPCR and other test are carried out to detect the disease. But they take somewhat longer time. So researchers are using artificial intelligence based techniques especially deep learning methods to develop new models using the CT scans (CTS) and chest X-ray (CXR) images of the patients to detect the disease in real time. This work focuses on the methods developed so far for detecting Covid-19 using convolutional neural network and compare their performances.
A Deep Learning Based Image Steganalysis Using Gray Level Co-Occurrence Matrix
A Deep Learning Based Image Steganalysis Using Gray Level Co-Occurrence Matrix, 2022
Image steganalysis is the technique to identify steganography in images and if possible try to pr... more Image steganalysis is the technique to identify steganography in images and if possible try to predict the quantity of hidden data. Targeted steganalysis need the knowledge of the steganographic algorithm used to embed the data, whereas blind steganalysis is independent of the embedding process. Its objective is to find patterns in the stego image that are generated due to steganographic process. This work proposed an elegant technique of blind steganalysis which takes input clean image from a benchmark dataset and find the co-occurrence matrix of grayscale image(GLCM) for four pixel pair direction and produces average GLCM. After that PCA and Haralick features are generated from average GLCM. Next, steganographic embedding is applied to the clean images with Steghide application with different payload. Each of this stego images are applied similar feature extraction process as clean images in the dataset. Now a deep neural network based model is trained on the prepared dataset with proper label. The proposed method gives 90.93% and 84.63% accuracy for LFW and BOSS dataset respectively and outperform many similar algorithms.
Invisible Watermark based Image Authentication System with 5/3 Integer Wavelet Transform, 2022
The authentication techniques deals with the originality of an object. It may be image, text, aud... more The authentication techniques deals with the originality of an object. It may be image, text, audio etc. Image authentication widely use watermarking techniques both in spatial and transform domain, especially discrete wavelet based techniques are preferred for features like multilevel analysis and lossless reconstruction. Again, integer wavelets has added advantage of generating only integer coefficients which further make computation simpler and faster. This work initially used forward 5/3 integer wavelet transform (IWT) on cover image to generate sub bands. A watermark image is taken and its SHA256 hash code is generated. The watermark and hash code are embedded in two different sub bands using dynamic random position map and the sub bands are inverse 5/3 IWT transformed to generate stego image. During extraction, the opposite process is adopted and the hash code of extracted watermark is computed and compared with original hash code for verification of authenticity. The experimental observation of the proposed method revealed around 61.5dB peak signal
to noise ratio (PSNR), near zero mean square error (MSE) and very high structural similarity index measure (SSIM) with 8448 bit payload and PSNR 55 dB with 33024 bits of payload.
Tropical Cyclone Classification using Deep Learning, 2022
Tropical cyclones (TC) are among one of the deadliest natural disasters which affect millions of ... more Tropical cyclones (TC) are among one of the deadliest natural disasters which affect millions of people living in coastal areas around the world. In early days limited tools were available to analyze the huge meteorogical data that were generated continuously over time. With the advent of computing power and artificial intelligence based techniques it is now possible to predict the origin, landfall and intensity of the tropical cyclone by collaborative efforts of the resources available in countries around the world. The real time data analysis plays a major role. From early simulation models built upon the hydrological and satellite data to current sophisticated data driven deep learning models are continuously evolving to serve the human civilization to combat cyclones by providing accurate early warning systems and making efficient disaster preparedness. This paper studies the deep learning based systems along with few early Mesoscale systems to predict TC and compared their relative performances.
International Journal of Computer Sciences and Engineering, 2022
Image steganography is used as a covert communication technique which hides secret data in cover ... more Image steganography is used as a covert communication technique which hides secret data in cover image intelligently so that it is visually imperceptible. This is often used by individual or organization with bad intent to harm people, organization or society. Steganalysis technique is used to break these systems to extract the secret information, reveal such covert communication and thwart imminent threat. Steganalytic techniques can be broadly classified as targeted or blind. In the former the knowledge of steganographic system used should be known and the latter adopts a more general approach where no knowledge of the process used to hide data is required. This paper studies some well-established statistical methods of targeted steganalysis and gray level co-occurrence matrix based blind steganalysis and compare their performances.
Bipartition Using Genetic Algorithm
Circuit partitioning is the most critical step in the physical design of various circuits in VLSI... more Circuit partitioning is the most critical step in the physical design of various circuits in VLSI design. In this paper, Genetic algorithm for circuit Bi-partitioning has been attempted. The genetic operators can easily be applied in this type of problem. In the partitioning main objective is to minimize the number of cuts. This chapter addresses the problem of partitioning and particular the use of the genetic algorithms for circuit partitioning. The objects to be partitioned in VLSI design are typically logic gates or instances of standard cell. Here the circuits are considered as graph where the nodes represent logic gates or cells and edges represent the connection between these gates or cells. Hence circuit partitioning problem becomes as graph partitioning problem. The algorithm can partition circuit into two sub-circuits. Our method calculates the fitness value and discards solution with low fitness value. The increase in number of crossover point does not necessarily increas...
Circuit Bipartition Using Simulated Annealing
Circuit partitioning is the most critical step in the physical design of various circuits in VLSI... more Circuit partitioning is the most critical step in the physical design of various circuits in VLSI design. In this paper, simulated annealing algorithm for circuit Bi-partitioning has been attempted. In the partitioning main objective is to minimize the number of cuts. This chapter addresses the problem of partitioning and particular the use of the simulated annealing algorithms for circuit partitioning. The objects to be partitioned in VLSI design are typically logic gates or instances of standard cell. Here the circuits are considered as graph where the nodes represent logic gates or cells and edges represent the connection between these gates or cells. Hence circuit partitioning problem becomes as graph partitioning problem. The algorithm can partition circuit into two sub-circuits. Our method calculates the fitness value and discards solution with low fitness value. The result of the simulated annealing is compared with Kernighan-Lin (KL) algorithm and the result is satisfactory.
IJRAR, 2022
In this modernistic chapter of information and technology, a galactic volume of data generations ... more In this modernistic chapter of information and technology, a galactic volume of data generations is happening every moment. Big data is a phrase that is referred to data sets that are not only big or, massive but also having velocity, variety, veracity, and value making it difficult to analyze, visualize, and store using conventional methods. Big data analytics provides an efficient method to analyze and store those varied and complex structured data for unfolding hidden patterns and correlates them. This work focuses on an overview of big data, big data sources, big data analytics, methods, and tools that are applied to big data, storing and managing big data, the diversity of big data in making decisions in different fields, advantages, previous works, future aspects, difficulties, and security issues.
Circuit Bipartition Using Genetic Algorithm
Circuit partitioning is the most critical step in the physical design of various circuits in VLSI... more Circuit partitioning is the most critical step in the physical design of various circuits in VLSI design. In this paper, Genetic algorithm for circuit Bi-partitioning has been attempted. The genetic operators can easily be applied in this type of problem. In the partitioning main objective is to minimize the number of cuts. This chapter addresses the problem of partitioning and particular the use of the genetic algorithms for circuit partitioning. The objects to be partitioned in VLSI design are typically logic gates or instances of standard cell. Here the circuits are considered as graph where the nodes represent logic gates or cells and edges represent the connection between these gates or cells. Hence circuit partitioning problem becomes as graph partitioning problem. The algorithm can partition circuit into two sub-circuits. Our method calculates the fitness value and discards solution with low fitness value. The increase in number of crossover point does not necessarily increas...
Automated System for Detection of White Blood Cells in Human Blood Sample
Determination of the WBC count of the body necessitates the detection of white blood cells (leuko... more Determination of the WBC count of the body necessitates the detection of white blood cells (leukocytes). During an annual physical checkup, generally doctors prescribe for a complete blood count report. WBC count is required to determine the existence of disease for symptom like body aches, chills, fever, headaches, and many more. The existence of autoimmune diseases, immune deficiencies, blood disorders, and hidden infections within human body can also be alerted by the report of WBC count. The usefulness of chemotherapy or radiation treatment, especially for cancer patients, is also monitored by this report. This paper introduces an automated system to detect the white blood cell from the microscopic image of human blood sample using several image processing techniques.
Iraqi Journal of Science
Machine learning-based techniques are used widely for the classification of images into vari... more Machine learning-based techniques are used widely for the classification of images into various categories. The advancement of Convolutional Neural Network (CNN) affects the field of computer vision on a large scale. It has been applied to classify and localize objects in images. Among the fields of applications of CNN, it has been applied to understand huge unstructured astronomical data being collected every second. Galaxies have diverse and complex shapes and their morphology carries fundamental information about the whole universe. Studying these galaxies has been a tremendous task for the researchers around the world. Researchers have already applied some basic CNN models to predict the morphological classes of the galaxies. In this paper, a residual network (ResNet) model is applied for this purpose. The proposed methodology classified the galaxies depending on their shape into 37 different classes. The performance of the methodology was evaluated using the data set provi...
Off-line signature verification system using weighted complete bipartite graph
2014 2nd International Conference on Business and Information Management, 2014
Handwritten signature is being used in various applications on daily basis. The problem arises wh... more Handwritten signature is being used in various applications on daily basis. The problem arises when someone decides to imitate our signature and steal our identity. In different area these signature are verified by human being. But the outcomes depend on the ability of the person who performs this verification. Thus, there is a need for an automated system that can, with a great degree of certainty, identify the authorized person. This paper proposes a novel approach of signature verification. Firstly, the given signature is represented by a means of a graph, whose nodes and edges describe certain properties at sample points and relationship between points respectively. Then, graph matching techniques are introduced to compute the similarity of graph for the given signature with the graph information store in the database. This principle can be implemented off-line handwritten signature recognition systems. The system is compared with Global feature approach and gives satisfactory result.
JPEG Steganography and Steganalysis – A Review
Advances in Intelligent Systems and Computing, 2015
Steganography and steganalysis are important topics in information hiding. Steganography refers t... more Steganography and steganalysis are important topics in information hiding. Steganography refers to the technology of hiding data into digital media without making any visual distortion on the media. On the other hand steganalysis is the art of detecting the presence of steganography in the media. This paper provides a detailed survey on steganography and steganalysis for digital images, mainly covering the fundamental concepts, the progress of steganographic methods for images in JPEG format and the development of the corresponding steganalytic schemes. As a consequence, a comparative study is also done on the strength and weakness of these different methods.
Uploads
Papers by BIBEK R A N J A N GHOSH
Steganography comes into play when the medical image becomes a potential storage of patient information. Generally
steganographic techniques give higher importance to the perfect recovery of secret data than the actual reconstruction of the
original image. But in medical system the actual reconstruction of the original medical image is importance as this is the
primary source of disease understanding. Hence in this context reversible steganography plays an important role in
telemedicine field in storing, maintenance and communication of electronic patient record within the image itself, so that
where a doctor located at a geographically remote location can diagnose a patient on the basis of a medical image generated
elsewhere with confidence. In this paper we have studied some elegant techniques for reversible medical image
steganographic techniques and compare their performances.
to noise ratio (PSNR), near zero mean square error (MSE) and very high structural similarity index measure (SSIM) with 8448 bit payload and PSNR 55 dB with 33024 bits of payload.