In this digital era, most important thing is to deal with digital documents, organizations using handwritten documents for storing their information can use handwritten character recognition to convert this information into digital. Handwritten character recognition is currently getting the attention of researchers because of possible applications in assisting technology for blind and visually impaired users, human-robot interaction, automatic data entry for business documents, etc. It is very difficult to recognize Devanagari character due to presence of header line, characters that are attached together and similarity of shapes of multiple characters. In this project the recognition of handwritten Devanagari characters is done using deep convolutional neural networks (DCNN) which are one of the recent techniques adopted from the deep learning community. I experimented on the data set of DHCP provided by ISI. A layer-wise technique of DCNN has been employed that helped to achieve the highest recognition accuracy and also get a faster convergence rate. The results of layer-wise-trained DCNN are favourable in comparison with those achieved by a shallow technique of handcrafted features and standard DCNN.

Preparing routine using manual method in colleges and schools is very time consuming and tedious task. This ends up with various classes with same teachers having more than one class at a time. Due to manual approach, proper use of resources is neither effective nor efficient. To overcome all these problems, I made an automated system as automatic routine generator. The system takes various inputs like number of subjects, classes and their capacity, teachers, maximal lectures a teacher can conduct, topics to be covered in a week or a lecture. Depending upon these inputs it generates possible time tables satisfying hard constraints for working days of the week. The administration can make changes or updates if they want some modification on the generated routine

One of the major problems that exists among students after graduation is the lack of connectivity with their classmates, institution and other students as well. There are barely platforms where students both current and alumni can connect with each other for various purpose like asking for help, look up their projects and current engagement, even to organize get together. Alumni Connect plays an important role in this context. It is a web-based platform where an institution provides its students both alumni and current ones with the opportunity to connect them together. Institution imposes rule to sign up its students based on their status i.e, either current or alumni. Then, the students of particular batch can see all other classmates on their timeline after login. There is a search functionality, which they can apply to search other students based on their name and connect with the person of choice.

Traveling is one of the important means through which people can learn, explore and discover new things. Before traveling, there would be various queries going on in our mind about which place to visit, what would be the cost, how is the accommodation facility of that particular place? In order to solve these issues, this project entitled “Chatbot for Tours Agency Using Chatterbot” that uses the python library Chatterbot, has been designed. This project is a chat-bot built with tour packages to various mesmerizing destinations which will leave no stone unturned in filling people’s heart with excitement, energy and enthusiasm. It is an artificial intelligence-based system that keeps the tour related data into its database and thus selects the appropriate answers on the basis of the queries defined by the user. It is very hard to answer all queries made by the customers to the tours and travels company. They many need to hire an employee for answering the queries which indeed increases the operational cost. Hence, the main purpose of building “Chatbot for tours agency using Chatterbot”, automated chat application is to answer the various queries and save time by providing quick answers and thus decrease the operational cost for tours company.

Steganography is derived from the Greek word stenographic, which means covered writing. It is the science of secret communication. The goal of steganography is to hide the existence of the message from unauthorized party. The modern secure image steganography presents a task of transferring the embedded information to the destination without being detected by the attacker. The new security conception hides secret information within the LSB of image where a secret key encrypts the hidden information to protect it from unauthorized users. In general, in LSB methods, hidden information is stored into a specific position of LSB of image. For this reason, knowing the retrieval methods, anyone can extract the hidden information. In this paper, I purposed an image based steganography that Least Significant Bits (LSB) techniques and pseudo random encoding technique on images to enhance the security of the communication. A particular image is selected as the cover image and then, any secret message of any size is hidden into the image. Also, the hidden image can be decrypted using this project. Not only this, the project is able to study the change in image properties when exposed to two different algorithm; Bitwise and X-OR algorithms

The purpose of this project is to detect fake news headlines that are circulated on social medias and have created a bias amongst people for a specific group usually political . News contain headline, author and its body which are not enough to find out if the news is fake or real but on a labeled data set that have already determined which news is real and which one is fake can be used to train and help detect fake news .To actually get results the categorical data can be explicitly divided into Fake and real or false and true to select feature and classify to generate a final model that can detect fake news. The result of the analysis can be used to detect fake news with its probability of being true or false based on classifier (logistic regression and Linear SVM) from sklearn and train lots of data to produce more accurate result in the future in the field of fake news detection.

Sentiment Analysis is a method for judging somebody's sentiment or feeling with respect to a specific thing written in a piece of text. It is used to recognize and arrange the sentiments communicated in writings. The web-based social networking sites like twitter draws in a huge number of clients that are online for imparting their insights in the form of tweets or comments. The tweets can be then classified into positive, negative, or neutral. In the proposed work, logistic regression classification is used as a classifier and unigram as a feature vector.

The purpose of this project is to help applicants to classify jobs based on their resume. There are various job portals available on the Internet, which consist of thousands of job descriptions. Most of the job descriptions are unstructured and presented in unorganized manner and due to this a user face difficulty in finding relevant and matching jobs. There are web sites which only allow users to use keywords to search the jobs, which makes job searching tedious and blind task. This project uses algebraic information retrieval method of vector space model. The applicants need to upload their resume to the system in order to get the set of job descriptions that match to their resume. Cosine Similarity Algorithm is used for calculating similarity value between the job description and applicant’s resume. This approach will help the users to get the relevant job descriptions and can restrict his/her job searching to some top job description of his/her choice. For experimental purpose, I have taken job datasets from Kaggle websites containing job description from Google Inc. and used resume of file format .txt as a query.

Authentication is one of the significant issues in the era of the information system. Among other things, human face recognition is one of the known techniques which can be used for user authentication. The facial recognition system is a process of identifying or verifying a person from a certain frame of digital image or a video frame from a video source. Facial Recognition system works with multiple methods, but in general, it works by comparing selected facial structures from given image with faces within a database. It can be additionally described as a Biometric Security System based on the application that can uniquely identify or verify a person by analyzing the facial patterns of a certain person. This project involves extracting the face from a given frame, some image preprocessing (clipping, grayscale conversion) and then recognizing the person in this image via machine learning algorithms. This can be achieved by using the Principal Component Analysis algorithm. The Haar Cascade algorithm detects the face in a frame and then we use PCA to extract the Eigenvector and then finally create an Eigenspace. Then the new face will be scattered over the Eigenspace and the Euclidian distance is calculated to identify the person. This project would help organizations to implement a Door Access System that is secure and relevant than traditional systems. For experimental purpose, I have taken the pictures of my friends and my own to test the system and got the accuracy of 74%

Nowadays financial data are increasing in huge amount and it is hard to manage those data. We cannot study all those data thus, we go for different statistical approaches. It would be good if we could visualize those data by representing it in graphical way. In this web application user upload a CSV file of a specific format. The calculation is made on some economic based methods like return on investment, moving window and volatility calculation. These methods are effective in an analysis of profitability of a company. Return on investment gives the time period in which the investment will be returned. Likewise moving window helps to figure out the fluctuations in the time series data and volatility calculation will help to determine either an investment is riskier or not. This web application can be used as a tool for analyzing the profitability of a company. This will help an investor and a company for a buying and selling of a stock.

All the centralized transaction (central database) needs to be verified by a trusted third party organization. A digital currency also known as a peer-to-peer electronic cash would allow online payments and transactions to be directly sent from one party to another without going through a financial institution. The peer-to-peer electronic cash is possible due to the help of blockchain nodes. Blockchain is the core technology behind Bitcoin and other cryptocurrencies that has drawn a lot of attention over the past few years and even in the present global market scenario. It is a distributed database that allows direct transaction between two parties (senders and receivers) without the need of any central authority. It can act a constructive technology and also as a disruptive technology. Wallet generation is a system to give an overview of how the blockchain works and as a medium bitcoins are used to describe the concept of blockchain nodes

Context Based Spelling Correction is a web application that corrects the spelling errors in sentences. The current version of the system does not have grammar checking function, and the initial goal of this project was to develop a standalone spell checker. But there are other types of errors that spell checkers and grammar checkers cannot detect, such as “I love eating desert”, where “dessert” is the correct spelling for “sweet snacks” but “desert” was typed. This kind of error is called a real-word error. The main focus of this project is to looks at how a context sensitive spell checker can be built to detect real word errors by using the information probability approach. The basic framework and background necessary to develop a context-sensitive spell checker will be covered. Different approaches to context-sensitive spell checking have been investigated and the complete design and implementation of each approach will be presented

Vocabulary mismatch problem is a common phenomenon in the usage of natural languages for information retrieval. Systems that rely on matching user’s query and documents solely on the basis of vocabulary suffer from problems of synonymy and polysemy. Research shows that on average 80% of the time different people (experts in the same field) will name the same thing differently. In such scenario, information retrieval must move beyond keyword matching and find relevant documents based on the concept. Latent Semantic Indexing is one such method that identifies patterns in the relationships between the terms and the concepts contained in an unstructured collection of text. This project implemented Latent Semantic Indexing with TFIDF to retrieve relevant documents for a search query on a corpus of machine learning research papers from ArXiv.

Hike Management System is a web application which is made for DWIT Hiking Club. In DWIT Hiking Club, it conduct many hiking events for students where students can explore their knowledge and talents from new places and also help to explore the nature. The information of Hike attain by students are passed through emails where the club member have difficulty in searching hike route and more number of students apply for hike and the record is placed in excel, it become harsh to manage the count and the black list of the hikers. So, this project is targeted towards providing students/member with an easy to use portal where they can apply for hike and member can input the record efficiently and inconvenient manner. Further this portal makes it easier for the club member to keep a track of hikers and their respective work.

Food is the most important requirement of every living beings on earth. Proper amount of calories is needed for the proper functioning of the body but an excess of is not good. According to nutrition.gov, obsessions with food, body weight, and shape may be signs of an eating disorder. These disorders can affect a person’s physical and mental health; in some cases, they can be life-threatening. So, a balanced diet is necessary to be free from such illness and maintain a healthy lifestyle. “Food Recognition and its Calorie Measurement using Tensorflow Object Detection API” was designed with the purpose of giving people knowledge about what and how much they eat in order to maintain a healthy lifestyle. The system was developed using tensor flow for object detection and was successfully implemented using Python and Open Cv at back end with an accuracy of 50-60%. In this system the food area, size and volume will be used to calculate the calorie in an effective way.

Effective feedback, both positive and negative, is very helpful. Feedback is the valuable information that will be used to make important decisions. For top performing companies ‘continuous improvement’ is not just a showy catchphrase. It’s a true focus based on feedback from across the entire organization. Similarly, in Deerwalk Institute of Technology, the process of taking the feedback is done with the help of Google Form which is traditional approach as the information cannot be managed and maintained easily. In the process of requirement gathering and analysis it was found that Google Form is less interactive and the information access and management is tedious. In order to solve this problem, an interactive rating system can be built with which a poll can be created and rating can be done on the provided parameter to obtain the feedback. This system allows to create a poll for the events from which the feedback is to be taken. The user rates the event on parameter on which the rating is available. The use of the system will help in the management and analysis phase. The system has been implemented with two graphical user interfaces: one for admin (creating the poll and analysis) and other user (rating the poll) using Java Programming and Grails framework

Plagiarism is one of the issues occurring mostly in academic community. Plagiarism is the "wrongful appropriation" and "stealing and publication" of another author’s language, thoughts, ideas, or expressions and the representation of them as one's own original work. People plagiarize under various circumstances but it does not justify the doing; posing someone else’s work as yours is equivalent to theft. So, In order to find out the plagiarized document, plagiarism checker is used which checks each and every document and displays the percentage of similarities of the file content. This project makes use of Smith Waterman Algorithm which performs local sequence alignment; that is, for determining similar regions between two strings or nucleotide or protein sequences. Instead of looking at the total sequence, the Smith–Waterman algorithm compares segments of all possible lengths and optimizes the similarity measure. This plagiarism detection is useful for academic purposes in schools, universities to check the plagiarism between documents so that wrong people would not be praised and the correct ones would not be de moralized.

This project is about a web application that generates mails when a task is assigned to a certain individual. When given a set of inputs in the web application, the system then generates a formatted mail which is sent to the selected receiver. It can be further developed to send scheduled mails using triggers. This project is aimed at generating automated mails to assign tasks to individuals working in a specific company. In general perspective, one of the primary means by which colleagues in a company communicate is through emails so this system uses that mode to assign tasks. The procedure while building this system has been simple. Initially, the design part was made in HTML/CSS and then the mail generation was done using SMTP. Therefore, with an aim to make the mode of communication easier and more structural, this project has been developed. This product has been running in the target company currently. The results have been conducted with no issues and the system has been effectively implemented in the company.

Unemployment is a serious social and economic concern almost all over the world. It leads to many social ills. Unemployment is an issue that governments try to address. When unemployment rates fall many social ills also see a decline. When people are gainfully employed it leads to social and economic well-being in the country. In order to provide part time jobs to people so that people can utilize their free-time and earn some money I have developed a web based application where every individual have the equal opportunity to apply for the particular job. This web –based application helps people to utilize the free

Database is an important part in the development of web application. In this digital era, data is the valuable asset for researcher and organization for any kind of work. Working with database locally could be tired some job for students. The term database also usually implies a series of related properties: data sharing, data integration, data integrity, data security, data abstraction and data independence. Database Management System helps manage the database in the system like delete, download, view the table. In this project, the management of database is done using basic PHP. Main root server can be accessed from login page which then will list all the available database in that system where we can download, delete, view the table.

Images have become an important part of any website. According to HTTP Archive, images make about 54% of the total webpage's weight. The Websites that load faster have better user engagement. So, if we can reduce the image size without compromising its visual quality, then it will have positive impact on page load time. Image Compression is used to minimize the amount of memory needed to represent an image. In this project, K-means Clustering algorithm has been used for the image compression. It is an unsupervised algorithm that automatically detects the clusters in the image data and groups each cluster of data together. The clusters are iterated over to reduce the redundancy present in the image to obtain the compressed image. In this approach the size of the compressed image is M X N X 3, therefore the image size is reduced maintaining the visual similarity between the original and the compressed image. The images were uploaded in a static website and the website was hosted locally, the load time of the website was compared and was found that the load time of the website with the compressed images was reduced in comparison to the load time of the website with the original images. Hence, it was concluded that the load time of the website is effected with the size of the image.

Tourism is an important asset of Nepalese economy. To facilitate the growing number of tourists visiting Nepal every year with necessary information and assist them during their visit, the authors have come up with an idea to make the process easier. The main idea behind this project is to gather all the necessary information under the same roof i.e. in a single application. This application is supposed to help tourists have a better experience. The application is a kind of itinerary generator for tourists. The app features a set of routes according to unique categories which contains complete information like transportation, things to do, etc. This application will also have a route generator, which can be shared with other users along with review and rating system. The app will also have a translation feature which will help tourists to understand Nepali. This application provides useful and relevant information to tourists regarding the particular place they are visiting. Some of such information may include a description of the very place as well as other places nearby, map of the area, weather, alerts about events happening in the very place, useful phone numbers, etc. The application would use a geolocation service to help the users find their location, navigate their way and find nearest hotels, homestays, restaurants and other such places. The application would help the users find their location, navigate their way and find nearest hotels, homestays, restaurants and other such places. Along with information, the application would also include the capability of organizing tours of various places. A route, similar to a travel itinerary, would involve a set of places arranged in a sequence for the tourists to travel in a specified time. Including a set of such routes provided by the creators, the application would also allow the users to make their own routes as well as customize them according to their various requirements. The users would also be able to share their routes and experiences with other users in an online forum.