Internship in Research Lab
Instructions for Inhouse Internship
Internship in Research Lab (IIT/NIT/University Research Lab based on call for research by respective faculty)
Students, who are interested in research based internship are supposed to work on the research projects given by faculties of IIT/NIT/CHARUSAT University.
As an outcome of project, one successful publication of a research paper is compulsory.
Students have to report at college regularly under the supervision of respective faculty.
List of CHARUSAT Research Lab Projects will be intimated by 1st April, 2023.
Faculty Open Research project
Faculty Name: Dr. Dhaval Bhoi
Research Project Title:Create parallel corpus: English-Gujarati
Abstract: This project aims to construct a comprehensive parallel corpus for English-Gujarati translation, catering to the growing need for language resources in the field of natural language processing and machine translation. The development of such a corpus holds significant importance in facilitating research and advancements in various language-related tasks, including machine translation, cross-lingual information retrieval, and sentiment analysis.
Technology:Google Translate can be used initially as a reference or a starting point for collecting parallel data. It can be used to generate initial translations and then manually verify and correct the translations for accuracy and appropriateness. Additionally, researchers can explore methods such as back-translation, where they translate Gujarati sentences back to English using Google Translate and compare them with the original English sentences to identify potential translation errors.
Dataset: Dataset is needed to be created from scratch.
Outcome:·The resulting parallel corpus will be made freely available to the research community, fostering advancements in English-Gujarati translation technology and facilitating cross- lingual research endeavors. It is anticipated that this resource will serve as a valuable asset for researchers, developers, and practitioners, enabling them to train and evaluate machine translation systems, develop linguistic tools, and explore various applications in multilingual natural language processing.
Team Size: - 3
Registration Form Link(Google Form): forms.gle/N2XT6sAAz9Z6awSMA
Faculty Name: Dr. Dhaval Bhoi
Research Project Title: Image-to-Text Translation in Gujarati
Abstract: This project proposes the development of a novel system capable of
generating/reading and displaying Gujarati text from input images.
Technology: To display Gujarati content present in an image, you would need a combination of technologies that involve Optical Character Recognition (OCR) for text extraction from the image and Gujarati language processing for displaying and handling the extracted text.
Dataset: Dataset is needed to be created from scratch.
Outcome:·This project will help to display the Guajarati content present in the image.
Team Size: 3
Registration Form Link(Google Form): forms.gle/N2XT6sAAz9Z6awSMA
Faculty Name: Dr. Dhaval Bhoi
Research Project Title: Gujarati Text Sentiment Classification
Abstract: This project proposes a Gujarati Text Sentiment Classification system utilizing Natural Language Processing techniques. It aims to accurately classify sentiment polarity in Gujarati text, enabling effective analysis of sentiment in Gujarati language content.
Technology: To implement Gujarati text sentiment classification, NLP libraries (e.g., spaCy), deep learning frameworks (e.g., TensorFlow), pre-trained language models (e.g., BERT), Gujarati text data, language resources, and evaluation metrics will be used.
Dataset: Dataset is needed to be created from the scratch [For Gujarati Language]. For English language IMDB dataset will be used.
Outcome: The outcome of the Gujarati text sentiment classification system will be the ability to accurately classify the sentiment polarity [Binary Sentiment Classification].
Team Size: 3
Registration Form Link(Google Form): forms.gle/N2XT6sAAz9Z6awSMA
Faculty Name: Dr. Arpita Shah
Research Project Title: Centralized ElasticSearch-AWS
Abstract:The research to proposes the development of a cloud-based data processing system that can handle large-scale data processing while maintaining high performance, security, and reliability.
Technology:Java, Python, Bash
● Development Tools: IntelliJ IDEA, PyCharm, Git, Azure DevOps
● Deployment Tools: Docker, Kubernetes, Helm
● Hosting Environment: Azure Kubernetes Service (AKS), Azure Container Registry
(ACR)
● Database: Elasticsearch, Logstash, Kibana (ELK) stack
Dataset: -
Outcome:·to centralize Elasticsearch and related components across different regions to reduce costs
Team Size: 3
Registration Form Link(Google Form): forms.gle/N2XT6sAAz9Z6awSMA
Faculty Name: Dr. Arpita Shah
Research Project Title: Process Unstructured Content: Aws Textract
Abstract:Extracting insights from unstructured documents manually is a tedious process and requires lot of time and effort. Learn how to add intelligence to process unstructured documents and extract meaningful information using AWS AI services such as Amazon Textract and Amazon Comprehend. In this demo, we showcase how easy it is to process unstructured content like product reviews and extract specific details related to the customer requirements, including overall sentiment of the reviews using Amazon Textract and Amazon Comprehend.
Technology:AWS/An IDE/Java Development Environment
Dataset: -
Outcome:·to centralize Elasticsearch and related components across different regions to reduce costs
Team Size: 3
Registration Form Link(Google Form): forms.gle/N2XT6sAAz9Z6awSMA
Faculty Name: Dr. Mrugendra Rahevar
Project List: sites.google.com/charusat.ac.in/mrugendrarahevar/project/open-project
Note: Students can find details of project and registration form on above link
Faculty Name: Prof.Martin Parmar
Application1: Smart Irrigation System: Internet of things(IoT)
The system aims to optimize water usage, enhance crop yield, and improve overall agricultural productivity. The primary objective is to conserve water resources by minimizing wastage and ensuring precise irrigation. The smart irrigation system will utilize sensors and weather data to determine soil moisture levels, allowing farmers to apply water only when necessary. By reducing water usage and preventing over irrigation, the system aims to address water scarcity issues and promote sustainable farming practices.
Application2: Water Quality Monitoring using Water Drone/Floater
IoT drone is set up on river water and floating on water to take readings such as water temperature, pH and dissolved oxygen. The floater acts as a service provider that provides sensor data on demand. All collected information is processed and transmitted to the IoT controller. Here, we used a forward proxy so that the information can only be directed to a legitimate device. The controller acts as an IoT-edged node with sufficient processing capability to process data and make decisions.
Link:
https://forms.gle/GeR1k2F6vXVRjQBz9
Faculty Name: Dr. Sneha Padhiar
Project Title: STRENGTHENING WEB APPLICATION SECURITY THROUGH TECHNICAL MEASURES
Details: This research aimed to investigate the various types of malware that commonly attack web applications, the strengths and weaknesses of current malware detection techniques, and technical measures that can be implemented to strengthen web application security and prevent malware attacks.
Team Size: 4
Registration Form Link(Google Form): forms.gle/N2XT6sAAz9Z6awSMA
Faculty Name: Prof. Deep Kothadiya
Research Project Title: Emotion Recognition With XAI
Abstract: Emotion recognition plays a crucial role in human-computer interaction, affective computing, and various other domains where understanding human emotions is essential. In recent years, deep learning techniques have emerged as powerful tools for automating the process of emotion recognition from various modalities such as facial expressions, speech, text, and physiological signals. This paper presents an overview of the state-of-the-art in emotion recognition using deep learning approaches.
Technology: Deep Learning, Python
Dataset: Any emotion recognition dataset
Outcome: Research Article
Registration Form Link(Google Form): https://forms.gle/HoAzAhX65QpXqhxr9
Faculty Name: Prof.Trusha Patel
Project Type : Development
Project Title : In house Software advancement in diagnostic microbiology
Technology : any web technology (php or any other)
Database : db compatible with selected technology (mysql or any other)
Outcome : Web portal for CHARUSAT microbiology department
Faculty Name: Pof. Rikita Chokshi
Research Project: Image Deconvolution using machine learning and deep learning approaches
Abstract: The technique to enhance or recover the quality of degraded images. The process typically begins with the collection of a dataset containing pairs of degraded images and their corresponding high-quality versions. These paired images are essential for training a deep neural network, often a convolutional neural network (CNN), to learn the intricate mapping between degraded and pristine images. During training, the model refines its parameters by minimizing the difference between its predictions and the actual high-quality images. The success of the model relies on its ability to generalize well to new, unseen data, which is assessed through a validation dataset not used during training. Fine-tuning may be applied based on validation results to improve overall performance. Once trained, the deep learning model can be applied to new degraded images, restoring their quality and revealing details that were previously affected by factors such as noise, blurriness, or compression artifacts. This technology finds applications in various fields, including medical imaging, satellite imagery analysis, and improving the visual appeal of photographs. The deployment of these trained models allows for real-world image restoration tasks, contributing to advancements in visual data processing and analysis. Overall, image enhancement with deep learning represents a transformative approach to enhancing the quality and information content of images affected by various forms of degradation.
Technology: python
Dataset: GoPro/HIDE/ DPDD
Outcome: Implementation result comparison and research paper