Browse
Recent Submissions
Item FACULTY PUBLICATION MANAGEMENT SYSTEM(I.O.E. Pulchowk Campus, 2023-06) AWAL, RUJA; MAKA, SANTOSH; POKHREL, SURAJ; DHAKAL, SUYOGFaculty publication management system is an integrated web application software that allows administrators to effectively manage all the faculty publications from record keeping to publication data analytics and visualization. Faculty evaluation system is also a part of the system that allows administrators to evaluate faculty members, lecturers and professors. This report presents design and development of the systems. Due to absence of a unified faculty publication management system, research publications of faculty members were spread across various research publication sites like GoogleScholar, ReserachGate, IEEE, etc and it became a tedious task for administrators to keep track of publications and generate associated reports for administrative purposes. Our system scrapes all the relevant data from such websites and presents user with a unified system to add, edit, approve, search, sort and generate reports of such publications as per faculty members or departments. The system additionally allows a user to perform data analytics and visualizations with bar chart and pie chart of research publications across faculty members and even specific departments. Later, the system uses the publication data of faculty members along with other evaluation metrics like years of service and objective feedback form data provided by students for evaluation of the user. Also, title based clustering of faculty members help to create clusters of members with similar research areas.Item IOE APP(I.O.E. Pulchowk Campus, 2023-05) SHARMA, NISHA; TIMALSINA, SAGAR; PURI, SANDIP; DHUNGANA, UDESHYAThe Institute of Engineering (IOE) in Nepal is utilizing various platforms for providing services to its students and staff. IOE APP is being developed as a comprehensive platform with a user-friendly interface integrating all the services in a place. This platform aims to integrate existing services while providing additional features through the use of plugin-like technology. The IOE APP strives to provide seamless services to its users and enhance the efficiency and centralization of college activities. The Agile methodology approach has been adopted to ensure that the IOE APP meets the requirements of its target audience. Furthermore, the IOE APP’s microservice architecture allows for scalability and extensibility, which can pave the way for similar applications to be developed for other organizations. This framework represents a shift away from manual-dependent systems and towards a completely application-based service.Item A INTEGRATED EDUCATIONAL TOOL(I.O.E. Pulchowk Campus, 2023-05) SHRESHTA, NIRAJ; PANTHA, RUPAK RAJ; NAGARKOTI, SITAL; DHUNGANA, SMARANThis report presents a mobile application project, titled ”IOE Portals: An Integrated Educational Tools,” aimed at addressing the digitalization gap in educational sectors, particularly in academic settings such as universities. The project is in response to the challenges faced by students and teachers in accessing academic resources and keeping up with the ongoing happenings at the institution. The mobile application includes features such as routine, syllabus, library, notification, personalized content, and probable exam question. Additionally, a web app for admin is created to manage the application’s backend. The project’s scope includes utilizing the application as a platform for sharing and accessing digital educational resources, organizing schedules, and providing informative resources. The report includes the project’s problem statement, objectives, methodology, and results, including a screenshots of the application’s interface. Overall, the project aims to provide digital assistance for the welfare of educational institutions and improve academic performance by enhancing accessibility to educational resources.Item VIDEO UPSAMPLING OF CCTV FOOTAGES(I.O.E. Pulchowk Campus, 2023-05) ARYAL, NIKHIL; POKHREL, SANDESH; BHANDARI, SANJAY; PANGENI, SANTOSHThe idea of super resolution and image upsampling have taken the field of computer vision by storm. New methods to upsample a grainy and low resolution videos are now the new chase. Our research is focused on upsampling a CCTV video through the use of deep learning techniques. Video superresolution often show sub-par results because they tend to have more components to process than their image counterparts, namely temporal dimension apart from the usual spatial dimension. In this research, we have studied these components and developed a pipeline that effectively processes the spatio-temporal information through optical flow, backed up by novel deep learning based VSR practices such as feature alignment, aggregation and upsampling. We examined and improved the pipeline based on the BasicVSR architecture and developed a model of our own by introducing residual in residual dense blocks. The new model RD-BasicVSR, was successful in surpassing the results of BasicVSR in both PSNR and SSIM metrics at same experimental settings.Item EASY READ(I.O.E. Pulchowk Campus, 2023-05) D.C., NIKESH; PANDEY, RAVI; CHHETRY, ROHAN; BANSAL, YUKTAMost of the information we perceive is through our vision. Visual impairment can hamper day to day task performing capability of an individual and among one of such crucial task is learning. On this note, to provide an aid in independent learning for visually impaired person, the system has been developed. The system captures image of textual documents using a mobile application. As the image is captured by people who are visually challenged, the image is likely to have distortions such as shadows and uneven illuminations. The captured image is sent to a remote server for pre-processing, involving binarization, noise reduction, and layout analysis to make it suitable for OCR processing.The pre-processed image is fed to an OCR engine to obtain textual format. The extracted text is then used to generate speech as final output, enabling people to hear the text in the captured image.Item 3D RECONSTRUCTION BASED VIRTUAL TOUR(I.O.E. Pulchowk Campus, 2023-04) KOJU, BISHAD; JYAKHWA, GAURAV; NYOUPANE, KRITI; MANANDHAR, LUNANumber of research and several methods have been proposed for 3D reconstruction from 2D images. The first is a triangulation approach based on determining the same points in images taken from different angles to approximate a point cloud in 3D space and then reconstructing the mesh. This is purely a computation-based approach. Another approach is to redefine 3D reconstruction problems as recognition problems and use the existing knowledge about 3D space and projection to reconstruct, much like how humans do. This knowledge is approximated using deep learning models. However, in these approaches, the mesh reconstruction part is extremely expensive. This cost can be reduced by trying to reconstruct the view rather than trying to reconstruct the mesh. Neural Radiance Field (NeRF) has been used to generate novel views. NeRF represents a scene using a fullyconnected deep network, whose input is a spatial location and viewing direction and output is the volume density and view-dependent emitted radiance at that spatial location. We synthesize views by querying 5D coordinates along camera rays and use classic volume rendering techniques to project the output colors and densities into an image. In this project, we have used the latter approach.Item HOME SEWA : AT-DOOR SERVICE DELIVERY SYSTEM WITH RANK BASED MATCH MAKING(I.O.E. Pulchowk Campus, 2023-04) POKHAREL, ARPAN; LAMSAL, CHIRAG; BASYAL, BIBEK; KAFLEY, SAUGATWith the progress in technology and IT, lives have been easier. Regardless of the advancements, in the present context in Nepal, although tasks and services have been somehow available for us all, different problems have arisen like delay in services, various scams, untrustworthy service providers and much more. Home Sewa is an application to provide users an access to at-door service delivery from the expert professionals. This application is designed to serve as a platform to connect customers and provide them quality service from the professionals. On the demand of a service from a user, the matchmaking system matches the user with the appropriate service provider using Elo ranking algorithm that incorporates important parameters associated with the delivery of services by service provider such as location, elo-rating and time availability. Home Sewa uses collaborative filtering model that gives personalized recommendations based upon the usage behaviour of similar users. This platform aims to provide a convenient and seamless experience for the customers seeking expert services.Item GENERATING VIDEO PRESENTATION FROM ARTICLES(I.O.E. Pulchowk Campus, 2023-05) POKHREL, AAGAT; AGARWAL, ANISH; PURI, BIPIN; PATHAK, BIRAJ BIKRAMThis project proposes a method for generating video slide presentations from text articles. The proposed method involves text parsing, feature extraction, clustering, ranking, summarization, slide creation, speech synthesis, and video generation. The BART model finetuned on dataset is used for feature extraction and abstractive summarization. The K Medoids clustering algorithm is used for clustering the sentence features, and the KNN algorithm is used for ranking. The markdown syntax and MARP are used for slide creation, and the Azure cognitive speech services and FFMPEG are used for speech synthesis and video generation, respectively. The comparision is drawn among BART large and BART base models on the CNN/DM dataset. The results show that the BART-large model outperforms in terms of ROUGE scores and employing the further pipeline generates coherent and informative video slide presentations.Item SHRUTI - A NEPALI BOOK READER(I.O.E. Pulchowk Campus, 2023-05) PAUDEL, PRABIN; SHAH, RAHUL; G.C., RANJU; KHADKA, SUPRIYAThe use of audiobook technology in the classroom has long been a viable instructional intervention for struggling readers. Shruti, an AI-generated Nepali book reader, is an application that generates a voice for the book. It is a text-to-speech(TTS) system that takes an input book in a PDF format. The PDF is extracted to text using Optical Character Recognition(OCR) and sent to the text-to-speech pipeline. The speech synthesis acts in two phases: spectrogram generation and vocoder output. The text is extracted, preprocessed, tokenized and sent to the modified Tacotron2 model for generating Mel spectrograms. The output in the form of Mel spectrograms is sent to the HifiGAN vocoder, which produces the sound. The synthesized sample of speech attained a Mean Opinion Score of 4.04 on the basis of naturalness, when audio samples were subjected to 28 volunteers. This sound is post-processed as a final output. The model has been deployed and integrated with a mobile application.Item NEURAL AUDIO CODEC(I.O.E. Pulchowk Campus, 2023-04-30) BARAL, SUBODH; PANDEY, TAPENDRA; BURLAKOTI, ACHYUT; BARAL, SIJALNeural audio codecs that use end-to-end approaches have gained popularity due to their ability to learn efficient audio representations through data-driven methods, without relying on handcrafted signal processing components. This research paper evaluates the performance of Neural Audio Codec in comparison to traditional audio codecs Opus and EVS in terms of audio quality and efficiency. The study highlights the limitations of existing audio codecs in leveraging the abundant data available in the audio compression pipeline and proposes deep learning-based models as a potential solution. The paper reviews recent advancements in deep learning-based audio synthesis and representation learning and explores the potential of deep learning-based audio codecs in enhancing compression efficiency. The study also addresses the limitations of existing models, including slower training times and increased memory requirements, by releasing open-source code and pre-trained models for further research and improvement. Experimental results show that our approach has comparable performance to widely used commercial codec OPUS at low bitrate, and a slight drop in performance compared to current deep learning-based frameworks but at the expense of significant improvement in speed and memory requirements. We have released our code and pre-trained models at https://github.com/AchyutBurlakoti/Neural-Audio-Compression for further research and improvement.Item AAWAJ : AUGMENTATIVE COMMUNICATION SUPPORT FOR THE VOCALLY IMPAIRED USING NEPALI TEXT-TO-SPEECH(I.O.E. Pulchowk Campus, 2023-05) BASNET, MAUSAM; POUDEL, NISHAN; DAHAL, SAMPANNA; SUBEDI, SUKRITIAs of 2016, more than 147,000 of the Nepali population suffer from some variant of speech or hearing impairments. Similarly, there has been a visible scarcity of reliable Text-to- Speech (TTS) engines in the context of the Nepali language. Aawaj is a dedicated mobile application that addresses these impediments to create augmentative communication support for the vocally impaired populace of Nepal using a Nepali TTS engine. BIt utilizes vocal features such as timbre, prosody, rhythm, etc., to create a natural-sounding TTS engine, based on the open-source Tacotron2 TTS architecture published by Google. Rare conditions such as cerebral palsy, spinal cord injury, muscular dystrophy, and amyotrophic lateral sclerosis (ALS) have also led to a physical impediment in speech generation for a large population. This report further proposes an Augmentative and Alternative Communication (AAC) platform using accessibility features such as text prompt generation that provides accessibility to the intended populace of this mobile application.Item AGRO-TECH(I.O.E. Pulchowk Campus, 2023-05) PANTHI, AKASH; ARYAL, AMRIT; KOIRALA, BIGYAN; KHANAL, BIPINThe mid-term report for our project titled ”Agro-Tech” provides an update on our progress and outlines the remaining tasks. Our platform aims to connect retailers and farmers for the efficient buying and selling of vegetables, with a focus on reducing post-harvest loss and ensuring food safety and freshness.We have made significant strides in developing the platform, including creating a user-friendly interface for retailers and farmers, and establishing partnerships with local farmers. We have also conducted market research and identified key areas for improvement, such as increasing the variety of vegetables available and expanding our reach to more retailers and farmers.Moving forward, we plan to focus on increasing the number of farmers on our platform and improving the logistical system for transporting vegetables from the farm to the retailer. We will also continue to prioritize food safety and freshness through the use of cooling infrastructure and adherence to global logistical standards. Overall, we are confident that our project has the potential to make a positive impact on the agro-based industry in Nepal, contributing to the overall development of the economy, reducing poverty and unemployment, and promoting equalityItem Nepali OCR(I.O.E. Pulchowk Campus, 2023-05) Bhusal, Abish; Chhetri, Gopal Baidawar; Bhattarai, Kiran; Pandey, Manjeethandwritten Nepali texts. Our system’s architecture consists of a Convolutional Neural Network (CNN) for feature extraction and a Recurrent Neural Network (RNN) for sequence recognition. The training dataset consists of around 80,000 Nepali handwritten images, which are preprocessed and augmented to increase the system’s robustness. NepaliOCR can be used to convert printed or handwritten Nepali text into editable digital formats, making it useful for a range of applications such as document digitization, language learning, and natural language processing. This paper presents an overview of the NepaliOCR system, including its architecture, training methodology, and performance evaluation. The lack of a reliable tool for Nepali handwriting recognition motivated us to develop this system. The system is developed as a part of the Bachelor in Computer Engineering Major Project. Machine learning has been used in the system to overcome the limitations of traditional computer systems. Our attempt to develop a tool for Nepali Handwriting Recognition using Machine Learning is discussed in this report. With the recent advancement in machine learning, our system shows great potential for practical applications in the digitization of Nepali handwriting.Item INFORMATION EXTRACTION FROM STRUCTURED DOCUMENT(I.O.E. Pulchowk Campus, 2023-04-30) KANU, AAYUSH SHAH; POKHREL, ADITHYA; BASHYAL, BISHAL; SHARMA, JANAKThis project proposes the use of the LayoutLMv2 model, a deep learning model, for information extraction from form-like documents. Specifically, the IRS 990 tax form was used as the dataset for testing and optimization. The information extraction process from form-like documents can be challenging due to the complex layout analysis and text recognition required to identify fields and corresponding values. The proposed model, LayoutLMv2, has demonstrated its effectiveness in these tasks, making it a promising solution for information extraction from form-like documents. The project resulted in the development of a web application and annotation tools that provide users with a user-friendly interface to upload documents and extract relevant information accurately and efficiently. The annotation tool enables users to label data and train custom models, while the web application streamlines document processing for businesses and organizations.Item “INFORMATION EXTRACTION FROM UNSTRUCTURED DATA”(I.O.E. Pulchowk Campus, 2023-04-30) LAMMICHHANE, AAYUSH; NEUPANE, AAYUSH; PAUDEL, ANKIT; LAMSAL, ASHISHIn today’s digital age, the digitization of paper documents like invoices and receipts has taken on more significance. Nevertheless, manually entering data from these papers can take a lot of time and be prone to mistakes, which causes inefficiencies and drives up expenses for enterprises. To solve this issue, we created a software platform that automates the process of collecting important data from scanned documents using deep learning technology, more specifically the LayoutLM architecture. Users can upload their scanned papers in bulk to our platform and choose which fields, including date, merchant name, and total amount, they want to extract. The technology is scalable and can manage high document volumes while preserving precision and effectiveness.The user-friendly interface makes it easy for users to upload and extract information from their scanned documents. Our platform offers significant benefits, including increased efficiency, accuracy, and cost savings, and has the potential to transform the way businesses handle physical documents. In this project, we will provide an overview of our software platform, including the technology behind it, its key features, and its potential applications.Item FRAMEWORK FOR DISTRIBUTED APPLICATION DEVELOPMENT(I.O.E. Pulchowk Campus, 2023-05) POKHAREL, PRANJAL; GHIMIRE, SANDESH; AMGAIN, SANSKAR; CHAD, TILAKWith the advancements in power, cost, and availability of microprocessors as well as the rapid growth in networking technologies and infrastructure, it is natural that we have developed technologies to allow hundreds of machines to connect with one another over a LAN for resource-sharing. These resources range from hardware devices to databases, files, computational power, and much more. With the benefits of such a system for developers and students in mind, this project, aims to be a framework using which resources of multiple systems within a computer network can be utilized. This framework is an attempt to abstract away the complexities of distributed application development by providing a platform to work upon. It is intended to be a robust, scalable, and safe platform for developing resourcesharing applications. The project itself is a culmination of our four years of engineering study, as we derive our knowledge base from teachings of the curriculum such as operating systems, distributed systems, data structures and algorithms, microprocessors, and computer networks. Thus, the project serves both facets of academia and practical implementation.Item ELECTRONIC MEDICAL RECORD STORAGE AND MANAGEMENT SYSTEM USING BLOCKCHAIN(I.O.E. Pulchowk Campus, 2022-05) KABDAR, PRABHAT KIRAN; THAKUR, PRIYA; KARKI, ROHAN; ARJYAL, SHREEMAbove all else, a person’s health is the most important thing, so any personnel’s medical records are unquestionably a very sensitive and priceless resource.If this knowledge is given to the wrong people, they may misuse it in unfathomable ways. The conventional form of paper file storage looks to be the most dangerous method of storage, although being used today. Additionally, computerized folder systems for storing records are not very trustworthy and may allow the relevant authority to abuse them. However, handling such data would have been made much simpler if the patient had been given complete control over who should have access to these sensitive information by cutting out the middleman. The blockchain concept is presented at this point since it has proven to be a great technology for decentralized data storage. Data security is best ensured by blockchain decentralized apps that link several entities, including as patients, doctors, and laboratories.Item IMPLEMENTATION OF CONSORTIUM BLOCKCHAIN FOR DECENTRALIZED KYC SHARING(I.O.E. Pulchowk Campus, 2023-05) ATREYA, MUKUL; ACHARYA, SANDEEP; CHAULAGAIN, SANGAM; TIWARI, SAUJANKnow Your Customer (KYC) process that is followed by the Financial Institutions (FIs) at present is highly inefficient and inconvenient for both FIs and customers. This process requires verification of customer’s identity documents independently by the businesses leading to high costs as well as wastage of resources. This project provides an efficient solution based on the Blockchain technology. The submission details of customers are collected only once for the verification process, irrespective of the number of financial institutions they register. The verified document is then shared with the organizations that require the information based on the customer’s approval. It leads to an efficient KYC process reducing costs and resources. This system uses private distributed file storage to store the identity documents and consortium blockchain technology is used to record and manage the KYC transactions ensuring security and transparency. The financial institutions act as the full nodes of the blockchain and the synchronization of blocks takes place through the P2P network.Item QUESTION SIMILARITY DETECTION AND ANALYSIS(I.O.E. Pulchowk Campus, 2023-04-30) SHRESTHA, MILAN; SHAKYA, NISCHAL; SWARNAKAR, NITESH; SUBEDI, ROSHANThe project aims to explore the effectiveness of using the SBERT model and vector database for performing question similarity analysis. The project involves building a vector database by training a sentence transformer model on a large corpus of text data. The vector dataset is then used to perform question similarity analysis by retrieving similar questions and similarity scores to a given search query. The model is trained on a large corpus of ALLNLI datasets, other paraphrase datasets such as MRPC, and PAWS, and the semantic similarity of datasets such as STS and finally adapted on 9,282 custom-prepared engineering datasets. The sentence transformer model is trained using the aforementioned datasets with MNR Loss as the loss function. The effectiveness of the model is evaluated by using the STS test dataset and test set of the MRPC. The result of the project demonstrates that using a sentence transformer model and vector database for question similarity analysis outperforms the baseline method of keyword matching. The approach achieved a spearman correlation value of 0.863 on the STS benchmark and an accuracy of 88.7% on the MRPC test. The Spearman correlation value in the SBERT paper for the NLI-large dataset was below 0.80. These values show that continuous training of the model on other datasets besides NLI helps to increase the performance and performs better for downstream tasks. This suggests that the use of the sentence transformer model and vector database is a promising approach for performing question similarity analysis, which could have significant implications for information retrieval systems.Item ANOMALY DETECTION IN SURVILLENCE VIDEOS(I.O.E. Pulchowk Campus, 2023-05) GURAGAIN, JIWAN PRASAD; SHRESTHA, KUSHAL; KUNWAR, LAXMAN; SUBEDI, YAMANAnomaly detection with weakly supervised video-level labels is typically formulated as a multiple instance learning (MIL) problem, in which we aim to identify snippets containing abnormal events, with each video represented as a bag of video snippets. Although current methods show effective detection performance, their recognition of the positive instances, i.e., rare abnormal snippets in the abnormal videos, is largely biased by the dominant negative instances, especially when the abnormal events are subtle anomalies that exhibit only small differences compared with normal events. This issue is exacerbated in many methods that ignore important video temporal dependencies. To address this issue, we use add information from the optical flow which captures the temporal relation between successive frames in a video. In this project, we explored the field of video anomaly detection and reviewed existing literature on the subject, as well as related topics such as action recognition and optical flow extraction.