This project focuses on developing an AI system for detecting cars using deep learning techniques, specifically transfer learning. The pre-trained Mask R-CNN ResNet-50 FPN model from the Detectron2 framework is utilized to fine-tune the object detection task on a custom dataset. The dataset, sourced from Kaggle, includes 100 images of cars, annotated using the Labelme tool and converted into COCO format. The process involves training the model on 80 images and testing it on 20 images to evaluate its performance. The goal is to leverage the knowledge embedded in the pre-trained model to enhance detection accuracy while minimizing training time and computational resources.Github