Basanta Joshi

• Doctor of Engineering
• Professor (Assistant) at Tribhuvan University

The integration of technology has revolutionized various industries of different domains, including agriculture. While agriculture is vital to many developing countries, reliance on traditional farming and the absence of data-driven methods have led to decreased productivity. Adopting new technologies is crucial for enhancing agricultural productiv...

Deep learning for sequence modeling has gained prominence during the past few years. To accomplish this, LSTM network architectures have shown to be quite helpful for forecasting the next output in a series. There are several ways to generate music and this paper presents a novel way to generate digital music. Leveraging the power of sequence model...

Named Entity Recognition (NER) is one of the vital task for many Natural Language Processing (NLP) tasks. In recent times, transformer architecture-based models have become very popular for NLP tasks including NER achieving state-of-the-art results. The Bidirectional Encoder Representations from Transformers (BERT) model especially has been found t...

Automatic speech recognition has allowed human beings to use their voices to speak with a computer interface. Nepali speech recognition involves conversion of Nepali language to corresponding text in Devanagari lipi. This work proposes a novel approach for developing Nepali Speech recognition model based using CNN-GRU. The data is collected from th...

Recently, there has been interest in generating natural looking handwritten characters that the observers could not distinguish whether they are generated or written by humans. The generation of handwritten characters is a popular research topic having several applications. In the context of Nepalese handwritten character, very few of the steps wer...

In the field of security analysis of an organization, identifying anomalous activities of user from log data for insider threat detection is difficult as well as important. Identification of such anomalous insider behavior is commonly achieved by use of behavior modeling. This paper presents an approach of one class learning, also known as unary cl...

With the advancement of technology, many things with greater accuracy and ease than past had been made possible. Using image processing and machine learning techniques, various noticeable achievements have been seen in medical science. This paper stands on the foundation of the Convolution Neural Network to diagnose the disease of patients from Che...

Timbre, Content, Rhythm and Prosody are four crucial aspects of speech. These aspects controls how person speaks. When two speaker utters same content, then other three aspects controls the differences in their speech. Converting any person voice to the targeted speaker voice by control over these three parameters is the the main work presented in...

Nepal suffers from earthquakes frequently as it lies in a highly earthquake prone region. The relief that is to be sent to the earthquake affected area requires rapid earliest assessment of the impact in the area. The number of damaged buildings provides us with the necessary information and can be used to assess the impact. Disaster damage assessm...

Mishra, SagarJoshi, BasantaPaudyal, RajendraChaulagain, DuryodhanShakya, SubarnaFacial emotion recognition has been an active research topic with its extensive applications in the field of health care, videoconferencing, security and authentication and many more. Human facial emotion recognition (anger, happiness, sadness, etc.) is still a challeng...

Huge loss in crop production occurs every year due to late identification of pant diseases in developing countries like Nepal. Timely and correct identification of such diseases with less dependency in related field expert can be more effective solution to the problem. Plants suffer from various diseases and correctly identifying them by observing...

Alumni tracking is a difficult task for any institution that has been running for a long time. The Department of Electronics and Computer Engineering (DOECE), Pulchowk Campus, IOE, has been offering undergraduate programs since 1994 A.D. and masters programs since 2001 A.D. The existing information about the alumni of the department is unmanaged an...

This research project is initiated with the aim to pursue study of intra- and inter-band frequency interferences due to allocation of frequencies for technologies like 5G. Specifically, this research will concentrate in identifying possible interference issues with already existing communication networks (in particular SRDs networks) while implemen...

Identifying anomalous motion behavior in video sequences is a challenging task. Manual annotation of a large number of surveillance videos is time-consuming because of the limited human brain's visual attention. This work presents a new framework to detect abnormalities from unlabeled videos using motion patterns for the normal and abnormal event....

A heterogeneous energy source, comprising more than one type of energy sources, can function as a reliable source of power. The constituent energy sources may have characteristics with varying degree of preferences. In this paper, a multi-agent based system is used in a heterogeneous power system to maximize the use of energy sources with preferred...

ANN is a computational model that is composed of several processing elements (neurons) that tries to solve a specific problem. Like the human brain, it provides the ability to learn from experiences without being explicitly programmed. This article is based on the implementation of artificial neural networks for logic gates. At first, the 3 layers...

Anomalies refer to any non-conforming patterns to the expected behavior in the system. The detection of anomaly in real time from logs arriving at very high velocity and are in huge volume requires a distributed framework with high through-put and low latency. In this research, statistical method has been implemented for finding the suspicious asso...

Coronary artery disease (CAD) has become very common nowadays and is the leading cause of death across the world. In this paper, an attempt has been made to develop a model for the classification of CAD by using clinical features, ECG, and features from laboratory tests of patients. Optimum features were selected with the help of domain expert advi...

The guided navigation has enabled users with minimal amount of training to navigate and perform flight mission of micro unmanned aerial vehicle (MAV). In non-urban areas, where there are no other aerial traffic and congestion, MAV takeoff & travel does not need much Global Positioning System (GPS) accuracy. The critical part seems to be during the...

The guided navigation has enabled users with minimal amount of training to navigate and perform flight mission of micro unmanned aerial vehicle (MAV). In non-urban areas, where there are no other aerial traffic and congestion, MAV take-off & travel does not need much Global Positioning System (GPS) accuracy. The critical part seems to be during the...

Communication is an important part of life. To use communication technology efficiently we need to know how to use them or how to instruct these devices to perform tasks. Automatic speech recognition plays an important role in interaction with the technology. Nepali speech recognition involves in conversion of Nepali speech to its correct Nepali tr...

Today data structuring, processing and transforming data from one system to another regardless of source is a big need. Need of big data was felt almost when first GPS technology was introduced. GPS technology has evolved significantly since big data technology became widely used. Big Data has long been the subject of interest for Computer Science...

Identifying anomalies from log data for insider threat detection is practically a very challenging task for security analysts. User behavior modeling is very important for the identification of these anomalies. This paper presents unsupervised user behavior modeling for anomaly detection. The proposed approach uses LSTM based Autoencoder to model u...

Automatic emotion recognition is a challenging task as human uses more than one varieties of modalities to express their emotion. The applications of emotional intelligence can be found in many domains including multimedia retrieval and human-computer interaction. With arrival of deep neural network, there is great success in determining emotional...

The project is initiated in accordance with the Contract Agreement between “The Water Resource Research and Development Centre (WRRDC)”, herein after referred to as the “WRRDC”, and “Laboratory for Information and Communication Technology, Information and Communication Technology Center, Institute of Engineering (IOE), Tribhuvan University (TU), he...

Surveillance cameras are widely being used in public places for security and monitoring purposes. Detecting abnormal motion pattern from surveillance video sequences is challenging task. Most of the existing methods are based on supervised technique. Supervised method groups feature points into normal and abnormal motion pattern using classifier. B...

Pattern recognition is very important for the identification of anomalous patterns in log messages. This paper presents pattern recognition in time series log data for anomaly detection. The proposed method uses Seasonal Auto Regression Integrated Moving Average (Seasonal ARIMA) to identify deviations between actual and predicted values. The deviat...

This paper presents a Neural Network based Nepali Speech Recognition model. RNN (Recurrent Neural Networks) is used for processing sequential audio data. CTC (Connectionist Temporal Classification) technique is applied allowing RNN to train over audio data. CTC is a probabilistic approach of maximizing the occurrence probability of the desired labe...

Image classification is a popular machine learning based applications of deep learning. Deep learning techniques are very popular because they can be effectively used in performing operations on image data in large-scale. In this paper CNN model was designed to better classify images. We make use of feature extraction part of inception v3 model for...

Analysis of log message is very important for the identification of a suspicious system and network activity. This analysis requires the correct extraction of variable entities. The variable entities are extracted by comparing the logs messages against the log patterns. Each of these log patterns can be represented in the form of a log signature. I...

Computer vision plays important role in Intelligent Transportation System (ITS) for traffic management and surveillance. This paper implements existing vision-based detection and tracking algorithms to detect and track motorcycles. While few research has been carried out for vehicle detection, but no research has been carried as far known for track...

I used LeNet, an architecture of convolutional neural network, to classify Nepali data-sets from 0 to 9. The number of images used to train was 2000 for training purposes and 200 for validation purposes. My network consists of 2 convolutional layers with 2 max pooling layers and RELU as my activation function with 2 dense layers and also used softm...

Object detection is the task of recognizing and localizing objects in an image. Object detection in images have many applications including object counting, Visual Search Engine, security, surveillance etc. Deep Learning based techniques for object detection are divided into two categories as region based approch and single shot approach. In this p...

The advent of internet and the growing number of digital media has increased the necessity of Music Information Retrieval systems within which Music Classification is a prominent task. Here, we present methods to perform genre based classification over five different genre and mood based classification using a mood model that maps mood onto a two-d...

This research provides a statistical method for detecting anomaly using multivariate data from system metrics. The spectral anomaly detection technique has been implemented using PCA (Principal Component Analysis) as unsupervised method to detect anomaly from data consisting of both normal and abnormal events. A comparison has been done with anothe...

Mining patterns from the log messages is valuable for real-time analysis and detecting faults, anomaly and security threats. A data-streaming algorithm with an efficient pattern finding approach is more practical way to classify these ubiquitous logs. Thus, in this paper the authors propose a novel online approach for finding patterns in log data s...

Log files consists of several independent lines of text data and contains information about the events from one or different services which may come from one or more nodes on the network. Mining patterns from these log messages are valuable for real-time analysis of network behavior and then detecting fault, anomaly and security threats. A sequenti...

In the new trend of 3D image velocimetry techniques for flow measurement, particle streak visualization is attempted in order to apply it to tomographic reconstruction of 3D particle streaks. The reconstructed particle streaks are then morphologically analyzed and their locations and three vector components are quantified. The velocity recovery res...

Novel 3D image analysis and particle matching techniques for the use in the volumetric particle tracking velocimetry have been developed and tested by using synthetic images and experimental images of unsteady 3D flows. A tomography based particle reconstruction scheme along with the subsequent process of individual particle detection and validatio...

http://repository.tudelft.nl/assets/uuid:b472e0a1-89a8-41b0-a8c8-b0bb0a94c916/A160_paper.pdf

Camera calibration is just an entry process but eventually an essential part of every particle image based velocity measurement technique. In the present paper, three representative camera calibration models are tested and their accuracy is compared for the use in the 3D particle tracking velocimetry (PTV). The three camera calibration models, prop...

Novel methodology of 3D particle tracking velocimetry (PTV) has been developed for the use in full 3D flow diagnosis experiments. The new PTV method makes use of an algebraic tomographic reconstruction technique for estimating the 3D locations of seeding particles in the measurement volume which is illuminated by a volumetric light source. After de...

New algorithms of 3D particle tracking velocimetry (3D PTV) based on a tomographic reconstruction approach have been developed and tested by using synthetic images of unsteady 3D flows. The new algorithms are considered not only in the tomographic reconstruction process of the fluid volume with particles but also in the subsequent process of indivi...

A self-organizing map (SOM) based algorithm has been developed for 3-D particle tracking velocimetry (3-D PTV) in stereoscopic particle pairing process. In this process every particle image in the left-camera frame should be paired with the most probably correct partner in the right-camera frame or vice versa for evaluating the exact coordinate. In...

Digital holography appears to be the next-generation technology for holographic diagnostics of particle fields and has been applied to holographic particle tracking flow measurements. In the digital holography, the Fresnel or Fourier holograms are recorded directly by the CCD or CMOS cameras and stored digitally. No film material involving wet-chem...