CN-121999470-A - Method for detecting fatigue state of driver based on generated type countermeasure network model

CN121999470ACN 121999470 ACN121999470 ACN 121999470ACN-121999470-A

Abstract

The application is suitable for the technical field of fatigue driving detection, and provides a method for detecting the fatigue state of a driver based on a generated type countermeasure network model, the method comprises the steps of firstly obtaining real-time face image information of a target driver, and then determining fatigue detection result information based on a preset driver fatigue detection model and the real-time face image information. Aiming at the problems of large noise of a camera image, unbalanced data type samples, poor space-time feature extraction capability of the existing algorithm and the like in a natural driving scene, the application provides a deep learning network based on an antagonistic network model, which aims to fully extract facial features of a driver under a shielding condition and accurately identify the fatigue state of the driver.

Inventors

LUO FANG
CUI YUAN
GUO XIAOTONG
ZHANG QIMING
LIU ZHONGZHENG

Assignees

清远职业技术学院

Dates

Publication Date: 20260508
Application Date: 20251231

Claims (10)

1. A method for detecting a fatigue state of a driver based on a generated countermeasure network model, the method comprising: acquiring real-time face image information of a target driver; and determining fatigue detection result information based on a preset driver fatigue detection model and the real-time face image information.
2. The method of claim 1, wherein the neural network architecture of the driver fatigue detection model includes an encoder and a decoder, each of the encoder and the decoder including six identical layers, each of the encoder layers including a self-attention sub-layer and a feed-forward neural network sub-layer, each of the decoder layers including a self-attention sub-layer and a feed-forward neural network sub-layer, and an encoder-decoder attention sub-layer being further disposed between the encoder and the decoder, each of the patch size parameters corresponding to the generator of the driver fatigue detection model being set to 32.
3. The method of claim 2, wherein the driver fatigue detection model includes PatchGAN discriminator, the PatchGAN discriminator being composed of a convolution layer, the PatchGAN discriminator being configured to divide the image into JTT-sized matrices, then discriminate all Patches separately, and finally output true or false based on an average of discrimination results of all Patches.
4. A method according to claim 3, wherein the PatchGAN discriminator has a loss function of: ; The loss function corresponding to the whole network of the driver fatigue detection model is as follows: 。
5. the method according to claim 4, characterized in that after the acquisition of the real-time face image information of the target driver, the method further comprises: A preset 3 multiplied by 3 template window is adopted as a local texture characteristic operator, gray values corresponding to the central pixels of the real-time face image information are used as thresholds, and gray values corresponding to other pixels in the periphery of the central pixels are compared with gray values corresponding to the central pixels; If the gray values corresponding to other pixels in the periphery of the central pixel are larger than the gray values corresponding to the central pixel, the gray values corresponding to the other pixels are assigned to be 1, otherwise, the gray values corresponding to the other pixels are assigned to be 0; based on a preset first calculation function, determining local texture characteristic values according to the central pixel and the other pixels, wherein the first calculation function is as follows: 。
6. The method of claim 5, wherein the driver fatigue detection model further comprises a discriminator based on feature separation, the discriminator for determining authenticity of the input sample, the discriminator comprising an auxiliary classifier for classifying the input sample.
7. The method according to claim 1, wherein after the determining of the fatigue detection result information based on the preset driver fatigue detection model and the real-time face image information, the method further comprises: Determining precision rate information based on a preset precision rate calculation formula, wherein the precision rate calculation formula is as follows: ; determining recall information based on a preset recall calculation formula, wherein the recall calculation formula is as follows: ; determining average precision information based on a preset average precision calculation formula, wherein the average precision calculation formula is as follows: ; determining average precision mean value information based on a preset average precision mean value calculation formula, wherein the average precision mean value calculation formula is as follows: 。
8. A driver fatigue state detection system based on a generative countermeasure network model, the system comprising: the real-time face image information acquisition module is used for acquiring the real-time face image information of the target driver; the fatigue detection result information determining module is used for determining fatigue detection result information based on a preset driver fatigue detection model and the real-time face image information.
9. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method according to any one of claims 1 to 7.

Description

Method for detecting fatigue state of driver based on generated type countermeasure network model Technical Field The application relates to the technical field of fatigue driving detection, in particular to a method for detecting a fatigue state of a driver based on a generated type countermeasure network model. Background Since the twentieth century, the demand and the maintenance of automobiles in China have been rapidly increased along with the acceleration of the urban process. With the increase of the automobile conservation amount, the occurrence rate of traffic accidents is rising year by year. Fatigue driving refers to a phenomenon in which the alertness, attentiveness, reaction speed, judgment and decision ability of a driver are reduced when the driver drives for a long time or is not asleep enough. The occurrence of fatigue driving is often difficult to avoid due to the numerous factors that lead to fatigue and from person to person. At present, the variety and complexity of the actual driving environment affect the determination of the fatigue state, for example, the factors such as illumination variation and camera shake introduce noise, so a new fatigue state detection method is necessary to identify the fatigue state more accurately. Disclosure of Invention Based on the above, the embodiment of the application provides a method for detecting the fatigue state of a driver based on a generated type anti-network model, so as to solve the problem of low recognition precision of the fatigue state in the prior art. In a first aspect, an embodiment of the present application provides a method for detecting a fatigue state of a driver based on a generated type countermeasure network model, the method including: acquiring real-time face image information of a target driver; and determining fatigue detection result information based on a preset driver fatigue detection model and the real-time face image information. Compared with the prior art, the fatigue state detection method for the driver based on the generated anti-network model has the advantages that the terminal equipment can firstly acquire real-time face image information of a target driver and then determine fatigue detection result information based on the preset driver fatigue detection model and the real-time face image information, so that a new fatigue state detection method is provided to more accurately identify the fatigue state, and the problem of lower identification precision of the current fatigue state is solved to a certain extent. In a second aspect, an embodiment of the present application provides a driver fatigue state detection system based on a generated type countermeasure network model, the system including: the real-time face image information acquisition module is used for acquiring the real-time face image information of the target driver; the fatigue detection result information determining module is used for determining fatigue detection result information based on a preset driver fatigue detection model and the real-time face image information. In a third aspect, an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the method according to the first aspect as described above when the processor executes the computer program. In a fourth aspect, embodiments of the present application provide a computer readable storage medium storing a computer program which, when executed by a processor, implements the steps of the method of the first aspect described above. It will be appreciated that the advantages of the second to fourth aspects may be found in the relevant description of the first aspect and are not repeated here. Drawings In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings used in the description of the embodiments or the prior art will be briefly described below. FIG. 1 is a flow chart of a method for detecting fatigue status of a driver according to an embodiment of the present application; FIG. 2 is a first schematic diagram provided by an embodiment of the present application; FIG. 3 is a second schematic diagram provided by an embodiment of the present application; FIG. 4 is a third schematic diagram according to an embodiment of the present application; FIG. 5 is a fourth schematic diagram according to an embodiment of the present application; FIG. 6 is a fifth schematic diagram according to an embodiment of the present application; FIG. 7 is a sixth schematic diagram according to an embodiment of the present application; FIG. 8 is a seventh schematic diagram of an embodiment of the present application; FIG. 9 is an eighth schematic diagram provided by an embodiment of the present application; FIG. 10 is a ninth schematic diagram provided by an embodiment of the present app