In this paper, the novel two-dimensional (2-D) fast algorithm for realization of 4$times$4 forward integer transform in H.264 is proposed. Based on matrix operations with Kronecker product and direct sum, the efficient fast 2-D 4$times$4 forward integer transform can be derived from the proposed one-dimensional fast 4$times$4 forward integer transform through matrix decompositions. The proposed fast 2-D 4$times$4 forward integer transform design doesn't need transpose memory for direct parallel pipelined architecture. The fast 2-D 4$times$4 forward integer transform requires fewer latency delays than the state-of-the-art methods. With regular modularity, the proposed fast algorithm is suitable for VLSI implementation to achieve real-time H.264/advanced video coding (AVC) signal processing.

In this study, for the application of visible-light wearable eye trackers, a pupil tracking methodology based on deep-learning technology is developed. By applying deep-learning object detection technology based on the You Only Look Once (YOLO) model, the proposed pupil tracking method can effectively estimate and predict the center of the pupil in the visible-light mode. By using the developed YOLOv3-tiny-based model to test the pupil tracking performance, the detection accuracy is as high as 80%, and the recall rate is close to 83%. In addition, the average visible-light pupil tracking errors of the proposed YOLO-based deep-learning design are smaller than 2 pixels for the training mode and 5 pixels for the cross-person test, which are much smaller than those of the previous ellipse fitting design without using deep-learning technology under the same visible-light conditions. After the combination of calibration process, the average gaze tracking errors by the proposed YOLOv3-tiny-based pupil tracking models are smaller than 2.9 and 3.5 degrees at the training and testing modes, respectively, and the proposed visible-light wearable gaze tracking system performs up to 20 frames per second (FPS) on the GPU-based software embedded platform.

To exploit an image-based in-cabin monitoring system for driving behavior and in-vehicle occupants detections to improve driving safety, in this paper, by installing a fisheye-lens camera at the in-car roof center and by using RGB-format images as inputs, the YOLO-based deep learning models, including YOLOv3-tiny, YOLOv3-tiny-3I, YOLO-fastest, YOLO-fastest-xl, and YOLO-fastest-three scales, are studied to be candidate detectors. The proposed in-cabin monitoring design can detect the normal and distracted driving cases and in-vehicle occupants including back seat passengers and pet dogs. The experimental results show that the YOLO-fastest-three scales model performs the best metrics for F1-Score and mAP, which are 95.89% and 97.16%, respectively. The YOLO-fastest-xl model has the best metric for false negative rate (FNR), which is 2.63%. By the software realization, the proposed design executes up to 30 frames per second (FPS) with the GPU-based embedded device.

An intelligent video-based drowsy driver detection system, which is unaffected by various illuminations, is developed in this study. Even if a driver wears glasses, the proposed system detects the drowsy conditions effectively. By a near-infrared-ray (NIR) camera, the proposed system is divided into two cascaded computational procedures: the driver eyes detection and the drowsy driver detection. The average open/closed eyes detection rates without/with glasses are 94% and 78%, respectively, and the accuracy of the drowsy status detection is up to 91%. By implementing on the FPGA-based embedded platform, the processing speed with the 640×480 format video is up to 16 frames per second (fps) after software optimizations.