Problem Statement
- USA-based client is developing technology solutions for Dentistry.
- New product development with AI (Artificial Intelligence) using 2D and 3D images of dental structures of patients as well as their voice textures was under consideration with R&D in terms of implementation of various algorithms.
Solution
- Detection of anomalies in teeth (Panoramic images and bitewing images).
- Research on different Artificial Intelligence (AI) models and applications applied to finding the types of anomalies in the teeth. o Word2Vec 2.0 Deep learning model implementation for audio to text conversion.
- Applying STFT, DCT; calculating periodogram for calculating MFCC coefficients and forming the image out of it.
- Using MFCC images to train CNN from scratch for classification of sentiments of a person.
- Research on LSTM, FFT, DCT, Wavelet, Hasing method, Sequence to sequence autoencoder model, Hierarchical retrieval method based on Hash Table and dictionary for audio Fingerprinting, Glove Embedding and TF-IDF for NLP.
- Deep Canonical correlation Analysis for Multi model Sentiment analysis.
- Removed Protected Health Information (PHI) from medical data. o Exploration on Audio Fingerprinting to help with anomaly detection in teeth and sentiment analysis of the patients’ reviews.