Ssu-Yuan Chang | National Chiao Tung University (original) (raw)
Uploads
Papers by Ssu-Yuan Chang
Sensors
This paper proposes a deep learning-based mmWave radar and RGB camera sensor early fusion method ... more This paper proposes a deep learning-based mmWave radar and RGB camera sensor early fusion method for object detection and tracking and its embedded system realization for ADAS applications. The proposed system can be used not only in ADAS systems but also to be applied to smart Road Side Units (RSU) in transportation systems to monitor real-time traffic flow and warn road users of probable dangerous situations. As the signals of mmWave radar are less affected by bad weather and lighting such as cloudy, sunny, snowy, night-light, and rainy days, it can work efficiently in both normal and adverse conditions. Compared to using an RGB camera alone for object detection and tracking, the early fusion of the mmWave radar and RGB camera technology can make up for the poor performance of the RGB camera when it fails due to bad weather and/or lighting conditions. The proposed method combines the features of radar and RGB cameras and directly outputs the results from an end-to-end trained deep...
2018 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW), 2018
This paper proposes a tracking-by-detection method with a weighted scoring mechanism to associate... more This paper proposes a tracking-by-detection method with a weighted scoring mechanism to associate the trackers and the detection results for accurate tracking and counting vehicles in a surveillance application. For the vehicle detection, the proposed method uses our robust PVA-lite deep learning model to detect vehicles. The experimental results show that the proposed method can achieve more than 95% in the average counting accuracy.
This paper proposes a low-complex triangular road signs detection and recognition algorithm that ... more This paper proposes a low-complex triangular road signs detection and recognition algorithm that can be implemented on an embedded system for real-time applications and maintains decent detection and recognition accuracy under inclement weather conditions. The proposed method is composed of the shape detection to locate triangular road signs followed by the two feature extraction methods to focus on their different contents, and the descriptor construction method to eliminate the noise and make the system robust. This work is implemented on a desktop computer as well on an automotive-grade Freescale i.MX 6 embedded platform. Under a video resolution of 1280x720, the proposed system achieves 161 fps on the desktop computer and 17 fps on the Freescale i.MX6 embedded platform with an overall accuracy of 93.33%.
Sensors
This paper proposes a deep learning-based mmWave radar and RGB camera sensor early fusion method ... more This paper proposes a deep learning-based mmWave radar and RGB camera sensor early fusion method for object detection and tracking and its embedded system realization for ADAS applications. The proposed system can be used not only in ADAS systems but also to be applied to smart Road Side Units (RSU) in transportation systems to monitor real-time traffic flow and warn road users of probable dangerous situations. As the signals of mmWave radar are less affected by bad weather and lighting such as cloudy, sunny, snowy, night-light, and rainy days, it can work efficiently in both normal and adverse conditions. Compared to using an RGB camera alone for object detection and tracking, the early fusion of the mmWave radar and RGB camera technology can make up for the poor performance of the RGB camera when it fails due to bad weather and/or lighting conditions. The proposed method combines the features of radar and RGB cameras and directly outputs the results from an end-to-end trained deep...
2018 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW), 2018
This paper proposes a tracking-by-detection method with a weighted scoring mechanism to associate... more This paper proposes a tracking-by-detection method with a weighted scoring mechanism to associate the trackers and the detection results for accurate tracking and counting vehicles in a surveillance application. For the vehicle detection, the proposed method uses our robust PVA-lite deep learning model to detect vehicles. The experimental results show that the proposed method can achieve more than 95% in the average counting accuracy.
This paper proposes a low-complex triangular road signs detection and recognition algorithm that ... more This paper proposes a low-complex triangular road signs detection and recognition algorithm that can be implemented on an embedded system for real-time applications and maintains decent detection and recognition accuracy under inclement weather conditions. The proposed method is composed of the shape detection to locate triangular road signs followed by the two feature extraction methods to focus on their different contents, and the descriptor construction method to eliminate the noise and make the system robust. This work is implemented on a desktop computer as well on an automotive-grade Freescale i.MX 6 embedded platform. Under a video resolution of 1280x720, the proposed system achieves 161 fps on the desktop computer and 17 fps on the Freescale i.MX6 embedded platform with an overall accuracy of 93.33%.