<![CDATA[Artificial intelligence technology is advancing rapidly, one of which is through the NVIDIA Jetson Nano platform, capable of detecting objects using the SSD MobileNetV2 model. However, the use of this platform is still rare, and references regarding its performance in various lighting conditions and object densities are limited. This study aims to evaluate the performance of Jetson Nano in object detection, measure frame rate, and analyze the accuracy of SSD MobileNetV2 in bright, dim, crowded, and non-crowded conditions. Frame rate testing is conducted in real-time over 15 seconds, while accuracy is tested using a confusion matrix from 20 video frames. The results show improved performance with increased frame rate, especially in 10W power mode compared to 5W. Accuracy testing also provides data on precision, recall, and F1-Score under various conditions, including bright, dim, crowded, and non-crowded scenarios, offering insights into the factors affecting the performance of Jetson Nano and SSD MobileNetV2.]]>
Copyrights © 2024
