An NLP-driven face recognition system to improve Alzheimer’s patient engagement and care [MEMORY LANE]

Alzheimer’s disease is a brain disorder that gradually impairs memory and thinking abilities, making it difficult for patients to recognize friends, family, and relatives. As the disease progresses, it becomes increasingly difficult for patients to use existing applications designed to help them identify their loved ones. Most existing applications have complex user interfaces, lack accuracy, and demand significant computational resources, making them unsuitable for wider utility. To address this issue, an intelligent assistant application named ‘Memory Lane’ is proposed to support patients with Alzheimer’s disease and their carers. Memory Lane works in four phases: (1) An offline speech recognition system. (2) A transformer-based intent classification model. (3) A CNN architecture that generates face embedding to compare L2 distance with existing stored embeddings for face recognition. (4) A module that retrieves information from MongoDB and uses TTS for communication. Overall, it has the potential to be a valuable tool in assisting people with Alzheimer’s and their carers in improving their quality of life.