G S Srenath Kumar

Text Summarization Application

Unleash the power of text summarization and extract the essence in a single line..

Welcome to our Text Summarization Application, where we combine the power of modular coding, advanced machine learning models, and a user-friendly interface to deliver accurate and concise summaries from your text.

Our application follows a well-defined process to ensure reliable and high-quality results. It starts with data ingestion, where your input text is seamlessly integrated into the system. Next, the data undergoes rigorous validation to ensure its integrity and quality. Then, the information is transformed and preprocessed to optimize it for the summarization model.

We leverage the cutting-edge Google Pegasus Daily Mail model, trained on the SAMSum dataset. This dataset comprises approximately 16,000 messenger-like conversations that have been carefully annotated with summaries. It reflects real-life conversations, encompassing a wide range of styles and registers, including informal, semi-formal, and formal conversations.

Our model, trained on this diverse dataset, is designed to generate concise summaries that capture the essence of your conversations. The summaries are formulated in the third person, providing a comprehensive overview of the key topics discussed.To provide a seamless user experience, we have developed a user interface powered by Flask API. Simply input your text into the provided input box, and our application will process it through the summarization model. Within seconds, you will receive a well-crafted summary that distills the essential information from your text.

For deployment, our application is compatible with popular cloud platforms such as AWS and Azure. This allows you to seamlessly integrate our text summarization capabilities into your existing infrastructure, ensuring scalability and reliability.

Experience the power of our Text Summarization Application and simplify your content consumption. Let us handle the complex task of condensing information, while you focus on extracting insights and making informed decisions.

Close Project

Django Web Application

This project was done by for an internship at Qurinom Solutions.

The use of this project is to transfer data files and announcements with the company systems connected to the servers.
The features of this project are that the company user can register and login into the web application only with his company webmail id.
The employees can share the data with all the other employees with this application, and the uploaded file can be edited at any time.

Also, the employees can search and view profiles of other users. They can share their uploaded files with any of those users.The additional feature is where the employees can set their profile pictures. But in default, the profile pictures added are apple face emojis.The type of files, such as images, videos, text files and different kinds of programs like python code,java code etc., can also be uploaded.

Technologies used are Python,Bootstrap,Django,JavaScript.

Close Project

Agriculture_Healthy_Harvest

"Revolutionizing agriculture for healthier harvests with Agriculture Healthy Harvest.

Welcome to our AI-powered solution for precision farming! Our platform is designed to assist farmers in making informed decisions by leveraging the capabilities of artificial intelligence. With our advanced algorithms, we enable farmers to predict the most suitable crops to grow, recommend appropriate fertilizers, and detect crop diseases using a combination of uploaded images and key data features such as soil composition, weather conditions, and water content.

One of the key features of our solution is crop prediction. By simply uploading images of your crops and providing relevant data features, our powerful algorithms analyze the information and generate predictions on the most suitable crop for your specific farming conditions. This helps you optimize your crop selection and maximize yield.

In addition, our solution offers fertilizer recommendations based on the predicted crop and soil type. We take into account various factors such as nutrient requirements and soil composition to provide tailored fertilizer recommendations that enhance crop growth and productivity.

Crop disease detection is another critical aspect of our platform. By uploading images of affected plants, our AI algorithms analyze the images and identify potential diseases. Timely detection of crop diseases is crucial for effective management and control, allowing farmers to take proactive measures to protect their crops.

To further support farmers, we have integrated a shopping cart feature into our platform. Once you receive crop predictions, you can conveniently add the recommended seeds and tools to your cart. Our aim is to simplify the purchasing process by offering a seamless experience and competitive prices.

To help you gain insights into your farming practices, our platform provides statistical analysis of crop growth in relation to soil and weather conditions. By exploring the intuitive dashboard, you can access valuable information about crop performance, trends, and correlations, enabling you to make data-driven decisions to optimize your farming practices.

To get started, you need to ensure you have the necessary prerequisites installed, such as Python 3, Flask, OpenCV, and NumPy. You can then clone the repository from our GitHub page and install the required packages using the provided command. Running the application will allow you to start utilizing the platform and taking advantage of its various features.

Healthy Harvest Website

https://farmers-shopping.vercel.app/
Close Project

Recycling Classification

Reduce, Reuse, Recycle: Be the change for a sustainable future.

Our classification project aims to address the significant problem of waste management in our country. With the majority of waste ending up in landfills, various environmental issues arise, including the increase in landfills, eutrophication, consumption of toxic waste by animals, leachate, and the escalation of toxins, leading to land, water, and air pollution.

The dataset used for training and evaluation consists of a total of 22,564 images, which were used for training the model, and 2,513 images, which were reserved for testing. These images encompass a wide range of different types of waste items, providing a diverse and comprehensive representation of the waste materials encountered in real-world scenarios.

Initially, a Convolutional Neural Network (CNN) deep learning model was employed for the training process. However, the model's performance in terms of accuracy fell short of the desired level. To address this issue and improve the accuracy of the model, we incorporated the MobileV2Net architecture. This particular architecture has demonstrated effectiveness in various image classification tasks and is renowned for its efficiency in terms of computational resources and memory utilization.

Furthermore, in an effort to further enhance the performance and robustness of our model, we applied data augmentation techniques to the dataset. Data augmentation involves the application of diverse transformations to the existing images, such as rotation, scaling, and flipping. By augmenting the dataset in this manner, we aimed to increase the variability and diversity of the training samples, thereby improving the model's ability to generalize and make accurate predictions on unseen data.

By following this implementation approach, we aim to provide an efficient waste classification system that can contribute to better waste management practices, reduce environmental pollution, and promote sustainability.

Close Project

Slatify Model

Unlock the power of concise knowledge with Slatify, where YouTube subtitles transform into meaningful summaries, accelerating productivity and expanding possibilities.r

Slatify is an innovative tool designed to revolutionize how you interact with YouTube subtitles. With seamless integration with the YouTube API, Slatify extracts subtitle data and transforms it into concise, easily digestible summaries. Not only that, but you can effortlessly transport these summaries directly into your Notion projects, making organization and collaboration a breeze.

Our cutting-edge technology is powered by the Pegasus dataset, a vast collection of news interactions between individuals, with summaries provided in a separate column. Leveraging this extensive dataset, we have trained our model, T5 Small, a powerful text-to-text transformer, on millions of instances. The result is a highly accurate summarization model that can distill lengthy text into key points and essential information.

Slatify is the ultimate NLP project, combining the power of machine learning, natural language processing, and data summarization. With our state-of-the-art algorithms and rigorous training process, we ensure that our model delivers reliable and accurate summaries, enhancing your productivity and saving you valuable time.

But we don't stop there. Slatify is more than just a web application; it's a complete product designed with convenience in mind. We are actively developing a Google extension that allows you to summarize YouTube subtitles within a minute, right at your fingertips. Now, you can effortlessly extract the essence of any video and capture the core ideas in a fraction of the time.

Slatify is not just a technological achievement; it's a collaborative effort. Our team of dedicated individuals, fueled by a shared passion for innovation, has worked tirelessly to bring this project to life. With the collective expertise and creativity of our friends, we aim to provide you with a seamless and intuitive user experience.

Close Project

Virtual Mouse with Hand Gestures

Controlling the mouse with Hand Gestures in Open CV.

I had created an AI-based Mouse Controller. Here we will first detect the hand landmarks and then track and click based on these points. We will also apply smoothing techniques to make it more usable.

Close Project

Geological Encryption Key

To create an encrypted password from the current latitude and longitude details.

This code is deployed and added to the PyPI library(the online package downloader), which is used to create an encryption password with a given key for a data file or a real-time locking system.

And this code is used the same in another autogenerated vital device. The locking system will get unlocked when both devices are in the exact longitude and latitude. We can get access when both the encryption is at the precise location as the password is created with the latitude and longitude.

Close Project

Movie Recommendation System

Unlock a world of film magic with our cutting-edge movie recommendation system.

Welcome to our AI-powered movie recommendation system! Our project focuses on providing personalized movie recommendations based on the given movie name. Using the power of artificial intelligence, we have developed a sophisticated algorithm that analyzes the input movie and generates a list of related movies that are likely to align with your interests.

One of the key highlights of our project is the modular coding approach. We have designed the system using a modular structure, which offers several advantages. It allows for better organization and management of the codebase, making it easier to understand, maintain, and extend the functionality of the recommendation system. The modular coding approach also promotes code reusability, making it more efficient and scalable.

When you enter the name of a movie, our recommendation system goes to work, leveraging advanced algorithms to analyze various attributes such as genre, actors, directors, and plot. It then compares these attributes with our extensive movie database to identify similar movies that you might enjoy. This process is performed seamlessly in the backend, ensuring you receive accurate and relevant recommendations.

Our project is a testament to the power of AI in delivering personalized experiences. By tailoring movie suggestions to your specific preferences, we aim to enhance your movie-watching journey and make it more enjoyable. Whether you are looking for a thrilling action flick, a heartwarming romance, or a thought-provoking drama, our recommendation system will provide you with a curated selection of movies that resonate with your taste.

We take pride in the modular coding structure that underpins our project. It not only enables us to develop and refine the recommendation system efficiently but also ensures that the system remains adaptable and scalable for future enhancements and improvements.

Close Project

Blind Navigation System

The Blind Navigation System is a remarkable technology designed to assist individuals who are blind or visually impaired in navigating unfamiliar environments with greater ease and independence. By incorporating audible directions, haptic feedback, and other sensory cues, this system provides crucial assistance for users to orient themselves and move around safely.

To develop this system, we leverage various technologies such as Python, OpenCV, and other relevant tools. These technologies work together seamlessly to deliver a comprehensive solution that addresses the specific needs of individuals with visual impairments.

The system comprises multiple components, including the main Python file (main.py) responsible for coordinating the different functionalities. The navigation_direction.py file enables live tracking of the user's current location and provides directions towards the desired destination. The object_distance_detector.py file assists in detecting distances between the user, objects, and other individuals using the camera feed. Additionally, the tkinter_application.py file serves as the graphical user interface (GUI) for the application, enhancing user accessibility and control.

Key features of the Blind Navigation System include an overview that provides a high-level understanding of its components and functionalities. It also outlines the hardware and software requirements necessary for running the system effectively. The user interface is designed to be user-friendly, ensuring easy interaction and control. The system's navigation features offer detailed descriptions of how it facilitates safe and efficient movement through various environments. Safety and security considerations are also addressed to ensure user well-being and protection.

Close Project

Java Registration Form for Admission

Storing the details of the members in MySQL server with the JDBC Connection..

Languages used are Jakarta Server Pages (JSP), Jakarta Servlet, Java Database Connectivity (JDBC), MYSQL.

Close Project

Strong Random Password Generator and Manager

Generating password and stored in database with encrypted.

This is a random password generator that also stores the secret password to the file. It also helps to get the username and password when adding a website to the input. This was made with Tkinter in python for GUI Design Interface.

Close Project

Tableau IPL Data Analysis

Various factors affecting the IPL Statistics, such as Highest Run Scorers, Most Wickets Taken by a Bowler, No of Matches held at a Venue, etc.

Using Tableau, visualize and understand the IPL dataset fed to it and make the calculations on the floor to have predicted results and charts.

Analyzing and Graphical Representation of the Top 10 ten Players with Score, Batting Average, Wickets Taken, etc.

Team members:

Santhosh Dhoni

G.S.Srenath Kumar

Close Project