Amiya Kumar Tripathy
Don Bosco Institute of Technology, Mumbai
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Amiya Kumar Tripathy.
international joint conference on computer science and software engineering | 2016
Niketa Gandhi; Owaiz Petkar; Leisa Armstrong; Amiya Kumar Tripathy
Food production in India is largely dependent on cereal crops including rice, wheat and various pulses. The sustainability and productivity of rice growing areas is dependent on suitable climatic conditions. Variability in seasonal climate conditions can have detrimental effect, with incidents of drought reducing production. Developing better techniques to predict crop productivity in different climatic conditions can assist farmer and other stakeholders in better decision making in terms of agronomy and crop choice. Machine learning techniques can be used to improve prediction of crop yield under different climatic scenarios. This paper presents the review on use of such machine learning technique for Indian rice cropping areas. This paper discusses the experimental results obtained by applying SMO classifier using the WEKA tool on the dataset of 27 districts of Maharashtra state, India. The dataset considered for the rice crop yield prediction was sourced from publicly available Indian Government records. The parameters considered for the study were precipitation, minimum temperature, average temperature, maximum temperature and reference crop evapotranspiration, area, production and yield for the Kharif season (June to November) for the years 1998 to 2002. For the present study the mean absolute error (MAE), root mean squared error (RMSE), relative absolute error (RAE) and root relative squared error (RRSE) were calculated. The experimental results showed that the performance of other techniques on the same dataset was much better compared to SMO.
international conference on tools with artificial intelligence | 2015
Tm Shahriar Sazzad; Leisa Armstrong; Amiya Kumar Tripathy
Dramatic improvements have been made in the field of digital image processing especially for biomedical image analysis over the past decade. With the availability of modern digital scanners, histopathology slides can be easily stored in digitized color image format. Therefore, histopathology digitized images have become a popular data source for both computer vision and machine learning techniques. There are several computer aided algorithms currently available to assist pathology experts to carry out their routine examination for detecting various tissues such as ovarian cancer cells and ovarian reproductive tissues. Automated detection of ovarian reproductive tissues is one of the important diagnosis interests for pathologists these days. One of the popular diagnosis preferences to identify ovarian tissues is ultrasound scanner. However, due to different shape, size and color, identification of ovarian tissues is a challenging task for ultrasound scanners as it process gray scale images. At present, pathological microscopic manual analysis is considered the best laboratory analysis practice for ovarian tissue cells although it is time-consuming, laborious and prone to errors. An alternate option would be to analyze these ovarian tissues automatically using color digitized images acquired from microscopic slides. In this paper a fully automated detection approach for color digitized image acquired from microscopic slides is presented and analyzed. The proposed method was found to be faster in comparison to other approaches. The approach also is beneficial as experts will not need to tune processing parameters for new batches of images. Experimental results from an analysis of the proposed approach using batch processing of a large number of images indicated high degree of accuracy and performance compared to the manual microscopic analysis.
computer based medical systems | 2011
Rebeck Carvalho; Rahul Isola; Amiya Kumar Tripathy
Traditionally the enormous quantities of medical data are utilized only for clinical and short term use. MediQuery puts to use this vast storage of information so that diagnosis based on this historical data can be made. There are systems to predict diseases of the heart, brain and lungs based on past data collected from the patients. We focus on computing the probability of occurrence of a particular ailment from the medical data by mining it using a unique algorithm which increases accuracy of such diagnosis by combining Neural Networks, Bayesian Classification and Differential Diagnosis all integrated into one single approach. The system uses a Service Oriented Architecture (SOA) wherein the system components of diagnosis, information portal and other miscellaneous services provided are coupled.
2015 International Conference on Technologies for Sustainable Development (ICTSD) | 2015
Imran Ali Mirza; Amiya Kumar Tripathy; Sejal Chopra; Michelle D'Sa; Kartik Rajagopalan; Alson D'Souza; Nikhil Sharma
Improving the quality of life for the elderly and disabled people and giving them the proper care at the right time is one the most important roles that are to be performed by us being a responsible member of the society. Its not easy for the disabled and elderly people to maneuver a mechanical wheelchair, which many of them normally use for locomotion. Hence there is a need for designing a wheelchair that is intelligent and provides easy maneuverability. In this context, an attempt has been made to propose a thought controlled wheelchair, which uses the captured signals from the brain and eyes and processes it to control the wheelchair. Electroencephalography (EEG) technique deploys an electrode cap that is placed on the users scalp for the acquisition of the EEG signals which are captured and translated into movement commands by the arduino microcontroller which in turn move the wheelchair.
international conference on communication information computing technology | 2012
Amiya Kumar Tripathy; Nilakshi Joshi; Steffy Thomas; Shweta Shetty; Namitha Thomas
The World Wide Web plays an important role while searching for information in the data network. Users are constantly exposed to an ever-growing flood of information. A wrapper is an application which helps in searching for Search Results Records (SSR) from multiple search engines. This helps in making the search more efficient and reliable. VEDD wrapper extracts the relevant SRRs from three search engines by filtering out the noisy and redundant records. Finally the unique set of records is displayed in a common VEDD search result page. The extraction is performed using the concepts of Document Object Model (DOM) tree. The paper presents a series of data filters to detect and remove irrelevant data from the web page. The data filters will also be used to further improve the similarity check of data records. Also, visual cues from the underlying browser rendering engine is made use to locate and extract the relevant data region from the deep web by the keyword matching technique.
2015 International Conference on Technologies for Sustainable Development (ICTSD) | 2015
Nilakshi Joshi; Amiya Kumar Tripathy; Suryakant Sawant; Tanvi Patel; Sailee Waghmare; Blessy Clusher
The Traffic problems are a critical problem that influences the travel time of the vehicles. The critical road conditions affecting the smooth movement of vehicles increases the traffic issue also the unorganized traffic flow, no dividers, steep curves, etc. gives birth to the accidents. Lack of information of the shortest path affects the emergency vehicle routing. Although calamities, accidents or casualties are subject to factors that cannot be predetermined, an precautionary measures can play a crucial role to avoid the same. The aim of this work is to develop a GIS based system for efficient vehicle management. GPS on client side gets coordinates of the client and sends it to the web server. The Server handles the request and updates the database with the locations of all the vehicles. The information is plotted on a GIS MAP and viewed by the admin to guide the user. By using the information obtained from the client; client can be routed to the most efficient path, he can also be updated with the accident prone areas. In this system dynamic shortest path is used for efficient travelling. The algorithm used for shortest path is based on Dijkstra algorithm. This data can be further exported in the form of daily reports. Using Geo Fencing concepts, alerts are sent to the client, when a client diverts from his routed path. This paper explores the foundation of GIS-GPS, location based services and Geo Fencing.
international joint conference on computer science and software engineering | 2016
Tm Shahriar Sazzad; Leisa Armstrong; Amiya Kumar Tripathy
Microscopic biopsy slides are used in the pathology laboratory by experts for general routine examination process to analyze various types of tissues. The manual microscopic analysis using biopsy slides is considered as a most viable approach but requires substantial amount of time and has observation variation issues among experts especially for smaller types tissue analysis. Available existing imaging modalities especially ultrasound scanner is a common device to analyze various types of tissues and is mainly suitable for larger tissue analysis. Computerized automated approach could be a more viable option as smaller tissues can be analyzed with less effort and in a short period of time with an acceptable accuracy rate. In this paper a complete review has been carried out on existing available approaches and a new modified approach has been presented for type P63 histopathology color images using three different magnifications which indicates improved accuracy rate.
2015 International Conference on Technologies for Sustainable Development (ICTSD) | 2015
Uma Sahu; Amiya Kumar Tripathy; Apurva Chitnis; Karen Aubrey Corda; Sharon Rodrigues
More and more E-commerce Websites provide products with different prices which made it hard for consumers to find the products and services they want. In order to overcome this data overload, personalized recommendation engines are used to suggest products and to provide consumers with relevant data to help them decide which products to purchase. Recommendation engines are highly computational and hence ideal for the Hadoop Platform. This system aims at building a book recommendation engine which uses item or user based recommendation from Mahout for recommending books. It will analyze the data and give suggestions based on what similar users did and on the past transaction history of the user.
2015 International Conference on Technologies for Sustainable Development (ICTSD) | 2015
Amiya Kumar Tripathy; Dipti Jadhav; Steffi A. Barreto; Daphne Rasquinha; Sonia S. Mathew
Many ways of communications are used between human and computer, one of them includes using gesture which is considered to be one of the most natural ways in a virtual reality system. Because of its intuitiveness and its capability of helping the hearing impaired and speaking impaired. Hand gestures enable deaf people to communication during their daily lives rather than by speaking. A sign language is a language which, instead of using sound, uses visually transmitted gesture signs which simultaneously combine hand shapes, orientation and movement of the hands, arms, lip-patterns, body movements and facial expressions to express the speakers thoughts. in order to bridge this gap of sign to speech, Voice fOr the Mute (VOM) aims to develop a system that will take real time images and convert them to speech with text as an intermediate taking into considerations all the limitations observed by a 2D system we will be considering only fingering spelling in our system i.e. Take input in the form of finger-spelling of alphabetic signs and providing the resultant voice output. The system will be using a webcam for the input and processing of the signs will be done using Microsoft Visual Studio as an IDE and OpenCv modules. With this proposed system we aims to help the speech impaired community.
2011 Developments in E-systems Engineering | 2011
Rahul Isola; Rebeck Carvalho; Mangala Iyer; Amiya Kumar Tripathy
The amount of Medical data recorded in hospitals and its significance as an ever-growing source of information has been long known and proven. Though the importance of the information hidden in these records has never been doubted, this data has mostly been used only for clinical purposes. Only recently has this been properly mined for valuable information to be used for research and to develop systems that assist the medical fraternity. Mostly, the systems that make use of this information are domain specific systems that predict diseases restricted to their area of specialization (like heart, brain etc.). But these systems are limited and are not applicable to the whole medical dataset. Our system uses this vast storage of information so that diagnosis based on this historical data can be made. This system aids medical diagnosis in the whole dataset by computing the probability of occurrence of a particular ailment from the medical data. The system mines the data using a unique algorithm which increases accuracy of such diagnosis by combining Neural Networks and Differential Diagnosis all integrated into one single approach. The strengths of kNN, Hop field algorithm, SOM and P2P Grid Architecture are used to make the system unique and effectively enhanced.