Diabetic Retinopathy Detection Using OpenCV
‍
‍This project leverages computer vision to address a critical healthcare challenge: detecting diabetic retinopathy, a leading cause of blindness. Using OpenCV, the system preprocesses and classifies retinal images to determine disease presence.
‍
Key Features:
‍Dataset Utilization: Trained the model on train_1 images and tested its accuracy with test_1, ensuring reliable performance.
‍Innovative Workflow: Covered data preprocessing, model training, and validation, showcasing expertise in machine learning and computer vision.
‍
Achievement:
Secured 3rd place in the Science Project Expo at SRM University for its innovative approach and potential to enable early detection and treatment.

IoT-Based Room Vacancy Detection System

‍This innovative system utilizes Internet of Things (IoT) technology to monitor room occupancy efficiently and provide real-time feedback. It combines multiple sensors and visual/audio indicators to create a seamless and user-friendly experience.
‍
Key Components and Features:
‍Ultrasonic Sensor: Measures the distance of objects from the sensor, identifying if an object is close to the detection range. This ensures accurate monitoring of objects within the room.
‍PIR Sensor: Detects human presence based on infrared radiation, ensuring reliable identification of room occupancy.
‍UNO Board: Acts as the central processing unit, integrating input from sensors and controlling outputs effectively.
‍Visual and Audio Indicators:
‍LEDs: Provide clear visual signals, with a green LED indicating room vacancy and a red LED signaling occupancy.
‍Buzzer: Produces an audible alert when an object is detected at a close range, adding an extra layer of feedback.
‍
Functionality:
The system constantly monitors the room environment, gathering data from the ultrasonic and PIR sensors. This information is processed in real time by the UNO board to trigger the corresponding LED or buzzer response. The simplicity and efficiency of this system make it an ideal solution for applications requiring quick and accurate room occupancy monitoring.

Computational Analysis of Phrasal Verbs in Research Articles: A Corpus Study

‍This project explores the nuanced usage of phrasal verbs in academic writing, focusing on research articles from UK and Indian English corpora. By leveraging natural language processing (NLP) tools, the study aims to identify, predict, and analyze patterns in phrasal verb usage, shedding light on regional linguistic differences in scholarly communication.

Key Components and Methodology:
‍Phrasal Verb Identification with NLTK: Used the Natural Language Toolkit (NLTK) to systematically identify and predict the occurrences of phrasal verbs within research articles. This step involved tokenizing, tagging, and parsing sentences to extract verb-particle constructions, ensuring accuracy in identifying relevant phrases.
‍Linguistic Analysis with spaCy:Leveraged spaCy’s advanced NLP capabilities for a deeper linguistic analysis of the identified phrasal verbs. Through dependency parsing and part-of-speech tagging, patterns in verb usage were visualized and compared across the UK and Indian corpora.
‍Highlighting Regional Differences:The study focuses on examining how phrasal verbs are employed differently in UK English versus Indian English academic writing. It highlights variations in frequency, context, and stylistic preferences, providing insights into regional trends and their impact on scholarly expression.
‍
Outcomes and Significance:
‍Visualization of Usage Patterns: Created comprehensive visualizations that illustrate phrasal verb trends, offering an intuitive understanding of regional linguistic characteristics.
‍Insights into Academic Writing: Provided valuable findings on the stylistic and cultural influences shaping academic communication, potentially guiding researchers, educators, and students in improving writing practices.