Phonic Forge
Loading...
Date
Journal Title
Journal ISSN
Volume Title
Publisher
St. Francis Institute of Technology (SFIT)
DOI
Abstract
Phonic Forge is an innovative AI-powered platform designed to enhance therapy for stuttering, a disorder impacting millions globally by disrupting communication and quality of life. Traditional stuttering therapies rely heavily on manual observation, lacking the precision, adaptability, and scalability that artificial intelligence (AI) and machine learning (ML) can provide. This project leverages advanced deep learning techniques, including Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN) models, to accurately detect and classify stuttering behaviors in real time. By combining these models with large language models (LLMs), Phonic Forge enables personalized therapy sessions through conversational AI, creating a uniquely tailored and interactive therapeutic experience for users. The platform addresses critical gaps in data diversity by utilizing open-source datasets, such as UCLASS and SEP-28k, to train and test its AI models. This approach enables better generalization of the models, making them adaptable to various speech patterns and stuttering types, including complex and overlapping behaviors. Recognizing the importance of data privacy and security, Phonic Forge incorporates blockchain technology to create a decentralized data storage system. Given the sensitivity of speech data, the platform�s blockchain integration ensures confidentiality and security, minimizing risks associated with centralized storage by leveraging decentralized solutions like IPFS. This feature provides a crucial layer of trust, enhancing reliability in the platform. By using React for the frontend and FastAPI for the backend, the system ensures smooth, real-time interactions during therapy sessions, a key factor in effective engagement and user experience. Accessibility allows users, regardless of location, to experience a high-quality, technologically advanced therapeutic process that may not otherwise be readily available in their area. In addition to its technical innovations, Phonic Forge provides user-centric features for patients. The platform offers therapy progress analytics, allowing therapists to track improvements, identify areas for focused interventions, and adjust therapy plans as needed. This capability enhances therapeutic efficacy while encouraging users to actively participate in their journey, fostering motivation and engagement.
Description
Project Report submitted to St. Francis Institute of Technology (SFIT).
Citation
Stuttering Detection
Speech Therapy
Artificial Intelligence
Deep Learning
LSTM
Convolutional Neural Networks (CNN)
Large Language Models (LLMs)
Blockchain
IPFS
Conversational AI
Speech Therapy
Artificial Intelligence
Deep Learning
LSTM
Convolutional Neural Networks (CNN)
Large Language Models (LLMs)
Blockchain
IPFS
Conversational AI