This dynamic web application uses Flask to create a REST API connected to an SQLite database and React for the frontend development. It's fully containerized with Docker and orchestrated using Kubernetes, suitable for robust deployment. The project includes testing with PyTest, and a CI/CD pipeline with Jenkins.

This project creates an image caption generator that provides concise descriptions of images, aiding visually impaired individuals and facilitating multilingual communication. Trained on the Flickr8k dataset, the model uses a ResNext encoder and LSTM decoder, achieving strong results for specific image types. Future improvements include diverse datasets and attention mechanisms.

Investigating consumer-grade white pc-LEDs, and RGBW LEDs and modulation techniques such as OPPM, WDM, and VPPM for visible light communication in the dark to determine which luminaire-modulation scheme pairs offer the best throughput, link distance, quality of ambient illumination.