Integrated Model for High-Level Egocentric Video Understanding

This thesis examines how dashcam videos can be used to understand different road events and driving behaviors. It introduces practical solutions by developing an advanced AI model to analyze and describe these videos. The model combines visual and language information, allowing it to process and explain what is happening in the videos efficiently. Using cutting-edge AI technology, this model can accurately interpret and describe the events captured by dashcams. This research highlights how dashcam videos and intelligent AI systems can be used to enhance road safety.