In the rapidly evolving landscape of sentiment analysis, the integration of facial recognition technologies has proven to be an invaluable asset for understanding emotional responses. This project, "Sentiment Analysis by Facial Detection Using Haar Cascade and CNN," focuses on harnessing advanced machine learning techniques to accurately interpret human emotions through facial expressions. Central to this system is the Haar Cascade classifier for efficient and reliable face detection, which identifies faces within images and video streams. Coupled with a Convolutional Neural Network (CNN), the system effectively classifies emotions such as happiness, sadness, anger, and surprise, providing deep insights into user sentiments. The implementation of this framework enables real-time analysis of emotional states, facilitating applications in areas such as marketing, customer feedback, and mental health monitoring. The system outputs both visual representations of detected emotions and corresponding sentiment metrics, offering a comprehensive view of emotional dynamics without reliance on extensive datasets. This project aims to equip researchers and practitioners with a powerful tool to enhance understanding of human emotions through cutting-edge AI and machine learning technologies.
The proliferation of digital content has significantly increased the risk of plagiarism, creating challenges for maintaining academic integrity across various fields. This project introduces a robust document similarity analysis system designed to combat this issue effectively. Utilizing advanced text processing techniques such as tokenization, stemming, and shingle generation, the system efficiently identifies overlapping content between documents. Central to its functionality is the Rabin-Karp algorithm, which allows for precise string matching, enabling the detection of similarities even in paraphrased or restructured sentences. The user-friendly interface facilitates easy document uploads, providing detailed similarity reports that highlight matched content and offer insights into the degree of similarity detected. These reports serve as both a plagiarism detection tool and an educational resource, promoting proper citation practices and the importance of originality. By fostering a culture of ethical writing and originality, this system represents a significant advancement in the fight against plagiarism, empowering users to uphold academic integrity in an increasingly digital world.
Football Analysis System-This system automated player tracking and provides performance based metrics.
Sign Language Detection Using CNN- The need for accessible and reliable sign language detection has become increasingly important. Sign language is a critical mode of communication for the deaf and hard-of hearing communities, yet tools for real-time interpretation are limited and often inaccessible to the general public. This project aims to develop a software-based solution for sign language detection, using advanced image processing and computer vision techniques. Implemented in a Python-based Jupyter Notebook environment, the system uses hand tracking, gesture recognition, and machine learning algorithms such as Convolutional Neural Networks (CNN) to accurately identify and interpret sign language gestures.
In the evolving field of text recognition, enhancing the accuracy of Optical Character Recognition (OCR) systems is crucial for improving digital text processing. This project, "Improving Text Recognition in OCR using BERT and Levenshtein Distance," leverages advanced Natural Language Processing (NLP) techniques to refine text outputs. At the core of this system is the BERT (Bidirectional Encoder Representations from Transformers) model, which enables contextual word prediction to correct misrecognized text by understanding relationships between words within sentences. Additionally, the project employs Levenshtein Distance, a dynamic programming algorithm, to compute the minimal edit distance between predicted and actual text strings, ensuring accurate error correction. The system operates without the need for an external database, providing a streamlined approach to generating highly accurate text recognition outputs. This project aims to deliver a powerful tool for enhancing OCR systems, integrating sophisticated NLP models and dynamic algorithms to improve real-world text recognition applications.
Fake Nepali Currency Detection using ORB, SSIM and Contour Detection This project addresses the issue of counterfeit Nepali currency by developing a platform for reliable detection. The system utilizes advanced image processing techniques, including ORB for identifying and matching key points between images, SSIM for scoring the similarity of compared notes, and Contour Detection for feature extraction. Implemented in Python, the platform ensures efficient performance and reduces time complexity. By creating a custom, manually curated dataset, the system achieves improved accuracy, providing a dependable tool for detecting fake Nepali currency.
Suspicious email detection is a critical aspect of cybersecurity, aimed at identifying and filtering out unwanted or harmful emails, commonly known as spam. The goal is to protect users from phishing attacks, malware, and other malicious content.
Pneumonia being one of the most concurrent prevalent lethal diseases, is responsible for several thousand deaths every year. X-rays images are commonly used to make clinical diagnosis for pneumonia but are not as revealing as MRIs or CT scans and the diagnosis is prone to be erroneous. Trained models help to make an accurate and more reliable computer-aided diagnosis of Pneumonia based on the X-ray images of lungs. In order to evaluate the Sigmoid-function, which is a variant of Transfer Learning, a Convolutional Neural Network/ VGG19 layer is used to achieve good performance. The problems and potential enhancements with the model of Transfer Learning is also used and discussed. Pneumonia Detection application user able to input the X-ray images and able to view the result whether the patient has suffered from Pneumonia or not. It shows expected result for proper X-ray images within the scope of the system. Error messages were shown when required. Time taken was moderate for moderate sized images. Excessive delays were not encountered the final accuracy obtained was 86.388
Flight Fare Prediction: The price of airline tickets is changed dynamically depending upon the various features of the tickets like destination, source, time, number of stops. Due to ever changing fare of airline tickets, it is hard for customers to predict the flight fare before purchasing the ticket leaving them confused whether to buy the ticket or not. This project report presents a simple approach for the predicting the flight fare using various algorithms like Random Forest regressor, decision tree regressor, K- neighbor classifier, Linear Regression. The proposed model uses the datasets of Indian Airline Companies available in the web. The experimental results show that the proposed method achieves 0.79 score on R-squared test achieved by using Random Forest Regressor. This study demonstrates the potential of using Random Forest Regression for flight fare prediction.
Exam Surveillance System using Machine Learning An exam surveillance system encompasses the automation of proctoring procedures within universities or colleges, aiming to enhance the efficiency of highly educated staff. By implementing such a system, the objective is to dissuade students from engaging in cheating behaviors during exams. This is achieved through the automation of the identification of potential cheating events, utilizing detailed logs. The system streamlines the review process for proctors, reducing the reliance on manual examination and enabling a more efficient detection and handling of academic dishonesty. Moreover, extending its utility beyond examinations, the system can also be seamlessly applied to interview process, providing a secure and monitored environment for virtual interviews.
Deepfake detection using deeplearning- Deepfakes threaten digital trust and security. This project combines ResNext50 and LSTM to detect deepfakes with high accuracy, offering a futuristic solution for ensuring media authenticity in an AI-driven world.
Vehicle license plate detection using machine learning