Object detection in car cabin environment

• Main Author: Aarathy Ajay
• Other Authors: Yap Kim Hui
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; Engineering; Object detection
• Online Access: https://hdl.handle.net/10356/176691
• Institution: Nanyang Technological University

In the recent years, the field of Artificial Intelligence (AI) has increased dramatically. Most of the AV (autonomous vehicle) industry is focusing in enhancing safety features and comfort. Car cabin object detection is a critical step in the development of advanced driver assistance systems (ADAS) and self- driving cars. This report examines various object detection models, focusing on their performance in detecting objects in the car cabin environment. The study assesses the performance of these models in detecting common objects like passengers, pets, and items left on seats or floors. Additionally, the report investigates the impact of various datasets, model architectures, and training strategies on detection performance. Experimental results show that cutting-edge object detection models can accurately detect objects in the car cabin, highlighting their potential to improve safety and convenience in automotive applications. In this project, YOLO is used due to its high speed and accuracy to compare different type datasets and has shown a promising result. YOLO family used a Single neural network to perform object detection directly on images. To attain this goal, the dataset was carefully selected, and deep learning optimization techniques were implemented. This project firstly compares different YOLO models to check its performance and found that latest YOLOv8 has high performance as compared to other YOLO models. Then YOLOv8 used in different datasets to see its performance and found that dataset with both IR and RGB images gave high mAP.