Customized Vehicle Feature Control through Driver Recognition Technology

This paper presents a novel approach for customizing the vehicle feature through driver recognition technology. The system combines the Cultural Adaptive Face Recognition (CAFR) and Contrastive Language-Image Pretraining (CLIP) models with OpenCV to recognize drivers and monitor their driving behavior, it also provides customizing vehicle feature control. The system uses a face recognition module that includes vision models like CAFR, CLIP, and Qdrant vector search databases, it recognizes the driver of the car by comparing their face with a pre-existing database of faces using the CAFR model. The image and text description-based embeddings are generated using the CLIP model for the images from the CAFR model, enabling zero-shot face recognition. The semantic embeddings generated by CLIP are used by Qdrant to search for similar faces in the database. The paper also proposes a driver tracking module that utilizes video analysis to monitor driving behavior and detect anomalies. The system uses extracted driver information from the face recognition module to set up customized features for the driver, such as setting temperature, Seat adjustment and playlists, based on their preferences. Additionally, the system implements optical character recognition (OCR) to extract information from ID cards and other documents, further customizing features to the driver's needs. The system's novelty lies in its ability to integrate multiple technologies to provide a seamless and personalized driving experience, enhancing driver assistance. The paper's findings demonstrate the effectiveness of the system in recognizing and tracking driver behavior, as well as setting up customized features. This technology has the potential to improve road safety, reduce driver fatigue, and enhance the overall driving experience.