公开(公告)号：US20230154207A1

公开(公告)日：2023-05-18

申请号：US17043681

申请日：2020-08-18

Applicant: Nanjing University of Science and Technology

Inventor： Yong QI ,  Yuan ZHUANG

IPC: G06V20/59 ,  G06V20/40 ,  G06V10/82 ,  G06V10/77 ,  G06V10/776 ,  G08B21/18

CPC classification number: G06V20/597 ,  G06V10/82 ,  G06V10/776 ,  G06V10/7715 ,  G06V20/46 ,  G06V20/49 ,  G08B21/18

Abstract: A driver fatigue detection method based on combining a pseudo-three-dimensional (P3D) convolutional neural network (CNN) and an attention mechanism includes: 1) extracting a frame sequence from a video of a driver and processing the frame sequence; 2) performing spatiotemporal feature learning through a P3D convolution module; 3) constructing a P3D-Attention module, and applying attention on channels and a feature map through the attention mechanism; and 4) replacing a 3D global average pooling layer with a 2D global average pooling layer to obtain more expressive features, and performing a classification through a Softmax classification layer. By analyzing the yawning behavior, blinking and head characteristic movements, the yawning behavior is well distinguished from the talking behavior, and it is possible to effectively distinguish between the three states of alert state, low vigilant state and drowsy state, thus improving the predictive performance of fatigue driving behaviors.

公开(公告)号：US20250078516A1

公开(公告)日：2025-03-06

申请号：US18681766

申请日：2023-04-20

Applicant: NANJING UNIVERSITY OF SCIENCE AND TECHNOLOGY

Inventor： Yong QI ,  Xin ZENG ,  Yuan ZHUANG

IPC: G06V20/54 ,  G06V10/75 ,  G06V10/80 ,  G06V20/70 ,  G08G1/017

Abstract: The present disclosure discloses a method and a system for detecting an abnormal traffic behavior. The method of the present disclosure includes: retaining an abnormal static target vehicle in a traffic surveillance video in a background through background modeling; performing abnormal target detection, and obtaining a cropped picture of an abnormal target vehicle and a cropped video clip through cropping; performing anomaly start time estimation, inputting the cropped picture and the cropped video clip to a network model combining twin cross-correlation with pseudo three-dimensional (P3D)-Attention, labeling a classification label on the cropped video clip, and determining a video frame when abnormal behavior occurs; and determining whether a to-be-matched vehicle is an abnormal target vehicle, and determining a start time and an end time of abnormal traffic behavior with reference to the video frame that is obtained when the abnormal behavior occurs.