Driver attention area prediction method and system based on target dynamic information

Abstract

The invention discloses a driver attention area prediction method and system based on target dynamic information. The method comprises the steps that spatial features of video frame images and dynamicfeature maps of adjacent video frame images are extracted; important target screening is carried out on targets in the extracted video frame images, cross-scale fusion is carried out on target feature maps of different scales, and cross-scale target features are obtained; attention fusion is performed on the spatial features and the cross-scale target features, and a driver attention prediction network model is trained with the dynamic feature map; and the trained driver attention prediction network model is adopted to predict the driver attention area of the to-be-tested video frame image. Through an important target screening network, an important target possibly existing at the current moment is mined, and the important target is fused with the image spatial features to enrich the spatial expression ability of the model; the inter-frame dynamic information is extracted through the extraction of the dynamic feature map, so that the method is more sensitive to the motion informationof an important target, and the prediction precision of the attention of a driver is improved.

Description

technical field [0001] The invention relates to the technical field of visual salience area prediction, in particular to a method and system for predicting a driver's attention area based on target dynamic information. Background technique [0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art. [0003] Driver attention area prediction, also known as visual saliency prediction in traffic scenarios, is crucial for assisted driving and driverless driving in complex traffic scenarios. The selective attention mechanism of the human eye helps the driver to detect danger at the first time. If the driver concentrates on driving, he will respond quickly to avoid traffic accidents; but there are often drivers who are distracted and fatigue driving Therefore, for the assisted driving system, learning the experience of the driver's attention can help the driver discover the danger in ad...

AI Technical Summary

Problems solved by technology

[0004] However, the inventor believes that although the traditional in-car attention collection equipment can estimate the driver's attention based on the driver's eye movement, it is difficult to provide enough data; traditional machine learning methods and deep learning-based In the dynamic traffic scene prediction method of the driver's attention area, it is not very sensitive to the sudden appearance of vehicles, pedestrians and other potentially dangerous targets. When the phenomenon of vehicle overtaking occurs, it is still undetectable, that is, it cannot extract effective sports information;

[0005] Since the driving task is in a dynamic scene, when the ConvLSTM module is introduced for timing modeling, the motion information is inserted after the high-level semantics, and the low-level semantic features will have a lot of edge and color information, and may contain more motion information. However, adding The dual-stream network constructed by optical flow often causes the model to be too large due to too many network model parameters, which will affect the training and prediction of the network.

Images