Coe AI

About
​​​​​​​COE-AI

Skill Odisha collaboration with Tech Mahindra has set up a Center of Excellence in Artificial Intelligence. This lab is also used for taking up commercial activities for the government sectors, local industries of Odisha to help them harness the benefits of Al in their field of activity. This lab will also help to tap the potential of the students of Odisha to address the evolving technology needs of local industries, besides significantly improving their skills and future employability.

Phonetic Search

About Project

Phonetic search refers to a type of search algorithm that takes into account the sounds of words, rather than just their spellings, when searching for matching results.

Specifications

The frontend part of the this application is developed using HTML and CSS and the backend part is developed with Python and Django as database.

FAQ System

About Project

A FAQ (Frequently Asked Questions) system is a software application or feature that provides a collection of questions and answers related to a particular topic or product. The system is designed to help users quickly find answers to common questions without having to contact customer support.

Specifications

The frontend part of the this application is developed using HTML and CSS and the backend part is developed with Python and Django as database.

Sentiment Analysis

About Project

Sentiment analysis  is the process of using natural language processing (NLP) and machine learning techniques to identify and extract subjective information from text, such as opinions, emotions, and attitudes. The goal of sentiment analysis is to classify text into categories such as positive, negative, or neutral.

Text Generation

Text summarization using NLP (natural language processing) is the process of using machine learning algorithms to create a condensed version of a larger piece of text while preserving the most important information. The goal of text summarization is to provide a quick and accurate summary of the original text, which can be useful for tasks such as information retrieval and content summarization.

Text Summarization

About Project

Text summarization using NLP (natural language processing) is the process of using machine learning algorithms to create a condensed version of a larger piece of text while preserving the most important information. The goal of text summarization is to provide a quick and accurate summary of the original text, which can be useful for tasks such as information retrieval and content summarization.

Text generation using NLP (natural language processing) is the process of using machine learning algorithms to generate new text based on existing text data. The goal of text generation is to create coherent and meaningful sentences, paragraphs, or even entire articles that resemble human-written text.

KisanSure[Won 3rd Prize at Smart Odisha Hackathon’22]

About Project

PMFBY(Pradhan Mantri Fasal Bima Yojana) Loan disbursement scheme Automated process of ml techniques to calculate the claim percentage as well as the specified amount without intervention of human input. This seamlessly brings down the time of loan disbursement by over 37%.The system will streamline the loan disbursement process, reduce manual errors, and improve efficiency by leveraging advanced machine learning algorithms.

Text generation using NLP (natural language processing) is the process of using machine learning algorithms to generate new text based on existing text data. The goal of text generation is to create coherent and meaningful sentences, paragraphs, or even entire articles that resemble human-written text.

Specifications

  • ML Based Algorithm for insurance claim prediction based on basis of farmer’s input.
  • Also uses a different algorithm for deciding the exact number of days required to get the claim settled

MoSwasthya[Won 1st Prize at Smart Odisha Hackathon’22]

About Project

Bilingual AI Based App for Cardiac Risk Prediction for Rural Areas Based on Non-Medical Vital Inputs . The mobile app will be designed to collect and analyze non-medical vital inputs such as lifestyle habits, physical activity levels, diet, stress levels, and family history, among others. The app will also leverage machine learning algorithms to integrate user data with existing medical data and research on cardiac risk factors. The bilingual feature of the app will allow users to easily input data and receive feedback in their local language, overcoming language barriers that often hinder healthcare access in rural areas.The primary objective of the project is to provide a simple and accessible tool for rural populations to assess their cardiac risk and take preventive measures accordingly.

Specifications

The app uses ML models to learn the users’ medical history and research about cardiac risks. It uses HTML, CSS and JavaScript for the frontend of the website. 
Uses CNN based ensemble techniques for cardiac risk prediction from user based non medical inputs.
Uses over 4 lac data points for model training.

Profanity and PII Filter

Specifications

A profanity and PII (personally identifiable information) filter is a type of software application that is designed to automatically detect and filter out inappropriate language and personal information from text data. This type of filter is commonly used in social media platforms, online forums, and messaging applications to prevent the spread of offensive or sensitive content.

AI Based Avatar

About Project

The goal of this project is to develop an innovative and cutting-edge system that utilizes artificial intelligence (AI) to enable images to "speak" in a manner similar to how input videos do. This image-to-speech synthesis system will be designed to convert still images into realistic and dynamic video-like content with synchronized speech. The system will leverage state-of-the-art deep learning techniques such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to analyze the visual features of the input image and extract relevant information, such as facial expressions, gestures, and lip movements. These features will then be mapped to corresponding speech patterns using natural language processing (NLP) algorithms.

Specifications

Avatar uses ML model to lip sync the data. It uses Python libraries 
text-to-speech, speech-to-text and moviepy.  The backend uses Django libraries. Frontend is developed using HTML, CSS languages.

RAH(Rapid Access HealthCare)

About Project

This website is a user-friendly platform that helps individuals find hospitals and doctors located near them by using their IP or GPS location. With a simple interface and intuitive design, the website allows users to quickly and easily search for healthcare professionals in their vicinity.
  • The website provides information on healthcare providers, including user reviews and ratings.
  • Users can make informed decisions about which doctor or hospital to choose based on the information provided.
  • The website features doctors' schedules and areas of expertise, as well as additional services offered by the hospital.

Specifications

The frontend part is built using HTML, CSS and Javascript. While the backend is developed using Python.

Papers and Patent

Patent Name AI Based Emergency Healthcare Solution (India Patents Office No- 202331002146)

Specifications

This is an end to end application that detects health abnormalities based on user input through a smartphone or web based input
Involves all sorts of emergency based services like ambulances, blood banks etc.

Patent Name From Robots to Books: An Introduction to Smart Applications of AI in Education (AIEd) (Published in Cornell University’s Arxiv)

Specifications

Deals with a review of all the applications of AI in the educational sector ( pros and cons as well as their ethical implciations)
Some of them include Intelligent Tutoring Systems, Smart Assesments etc.

Paper Name An efficient prediction of diabetic from retinopathy using machine learning and signal processing approach (Published in IEEE)

About Paper

Novel Algorithm for diabetic retinopathy detection from medical images using Discrete Wavelet Transform (DWT)

Specifications

  • In this study, the retinal image is taken from a fundus camera of both healthy and diabetic retina.
  • Image pre-processing techniques, morphological operations are used to detect the statistical features and the histogram-based feature is extracted by using Discrete Wavelet Transform (DWT) which is the novel contribution of the proposed algorithm.

Paper Name Interpretation of Optimized Hyper Parameters in Associative Rule Learning using Eclat and Apriori ( Published in IEEE Xplore)

About Paper

​​​​​​​Hybrid Algorithm for detection of frequent items in Data Mining Based Operations

Paper Name Deep Learning Based Classification of the Big Four Snake Species Using Visual Features( To be Published in IEEE Xplore)

About Paper

Deep Learning based approach for identification of Big-4 Snake Species in Indian Subcontinent
CNN Based Approach to compare and evaluate various models for identification of Big 4 snake species

Paper Name Chaurah: Smart Raspberry-Pi Parking System ( Published in IEEE TechRxiv & ICCCT’23)

About Paper

An Intelligent Multilevel Parking System for Smart Cities
  • CNN(for OCR)
  • Raspberry Pi and Pi Camera for the camera module.
  • Car Plate extraction-> Number Recognition -> Database Storage

Paper Name Chaurah: Smart Raspberry-Pi Parking System ( Published in IEEE TechRxiv & ICCCT’23)

An Intelligent Multilevel Parking System for Smart Cities
  • CNN(for OCR)
  • Raspberry Pi and Pi Camera for the camera module.
  • Car Plate extraction-> Number Recognition -> Database Storage

COURSE MADE

COE- 401: Introduction to Deep Learning

  • Involved over 50 participants from various undergraduate backgrounds to get them into the world of Deep Learning.
  • Worked on over 10 hands-on projects
  • Consisting of over 12 theory sessions and 10 practicums.
  • Studied state of art CNNs, Generative Models, NLP, and Reinforcement Learning-based algorithms.
  • All participants submitted individual assignments, projects and worked on a final poster presentation.