Browsing by Author "Dr Wajahat Mehmood Qazi"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Automated Product Sensing and Sentiments
    (Library Information Services, CUI Lahore, 2023) Ahmad Fareed; FA18-BCS-151; Dr Wajahat Mehmood Qazi
    Our project, APSS, is a web application that analyses brands, competitors, hashtags and their market trends in real time using keywords. The user can input the keywords. The app uses Reddit and News APIs to get data about these keywords at the backend. We then pass this textual data from the VADER which is a pre trained model for text sentiment analysis. The model is available in NLTK package to help us in predicting the sentiments correctly. The VADER is better than TEXTBLOB in order to achieve the accuracy. Our project is a very effective monitoring tool which portrays the results in the form of mentions, graphs, percentages and comparison charts. It helps the new brands in market as well as established brands to analyze about the competitors in market and trends. This helps them devise their marketing strategies and improve their products accordingly.
  • No Thumbnail Available
    Item
    DEEPFAKE TEXT TO SPEECH (TTS)
    (Library Information Services, CUI Lahore, 2024) Muhammad Talha; FA20-BCS-007; Dr Wajahat Mehmood Qazi
    This project aims to develop an application that can generate voice clones based on provided text using a trained model, we named this app Deepcloning.ai. Many Text-to- Speech models exist, they are limited to specific voices. Our focus is on training a model with an English dataset to replicate a particular person's voice. Once trained, the model can produce Deepfakes voices for the provided text, offering advantages in education, cost- effective advertising, and the media industry. However, it's essential to acknowledge the potential for misuse, especially in serious crimes and impersonation, necessitating subsequent efforts for detection and prevention. This project's primary objective is Deepfakes text to speech creation, with ethical and security considerations in mind.