Method and device for detecting standardization of wearing mask

A computer device, comprises a memory, a processor and computer programs stored in the memory and able to be run by the processor, wherein the computer programs, when being executed by the processor, are configured to implement the operations of:

• receiving an image to be detected, wherein the image to be detected comprises an object to be detected;
• detecting the image to be detected based on a trained multi-task network model to obtain a region where the object wears a mask;
• comparing an area of the region where the object wears the mask to a predetermined threshold to determine a first detection result, wherein the first detection result is that the object is wearing the mask or that the object is not wearing the mask;
• in response to the object being determined as wearing the mask, determining comparison information of a mask area to be compared, and determining a target region in the image to be detected corresponding to the comparison information based on a face feature point template, wherein the comparison information represents information of a comparison portion to be detected when standardization of mask wearing is detected, and the face feature point template is determined based on a face region of the object to be detected;
• and
• obtaining a second detection result based on an overlapping degree between the target region and the region where the object wears the mask, wherein the second detection result includes the mask being worn standardly and the mask being worn non-standardly;
• wherein the determining comparison information of a mask area to be compared, and the determining a target region in the image to be detected corresponding to the comparison information based on a face feature point template, comprise:
• receiving input customized comparison criteria, generating the comparison information of the mask area to be compared based on the customized comparison criteria;
• and
• determining first feature point information corresponding to the comparison information;
• screening, from the face feature point template, feature points identical to the first feature point information;
• and forming the corresponding target region in the image to be detected based on the feature points;
• wherein the target region is a closed polygon formed by connecting the feature points.

A non-transitory computer-readable storage medium, having computer programs stored therein, wherein the computer programs, when being executed by the processor, are configured to implement the operations of:

• receiving an image to be detected, wherein the image to be detected comprises an object to be detected;
• detecting the image to be detected based on a trained multi-task network model to obtain a region where the object wears a mask;
• comparing an area of the region where the object wears the mask to a predetermined threshold to determine a first detection result, wherein the first detection result is that the object is wearing the mask or that the object is not wearing the mask;
• in response to the object being determined as wearing the mask, determining comparison information of a mask area to be compared, and determining a target region in the image to be detected corresponding to the comparison information based on a face feature point template, wherein the comparison information represents information of a comparison portion to be detected when standardization of mask wearing is detected, and the face feature point template is determined based on a face region of the object to be detected;
• and
• obtaining a second detection result based on an overlapping degree between the target region and the region where the object wears the mask, wherein the second detection result includes the mask being worn standardly and the mask being worn non-standardly;
• wherein the determining comparison information of a mask area to be compared, and the determining a target region in the image to be detected corresponding to the comparison information based on a face feature point template, comprise:
• receiving input customized comparison criteria, generating the comparison information of the mask area to be compared based on the customized comparison criteria;
• and
• determining first feature point information corresponding to the comparison information;
• screening, from the face feature point template, feature points identical to the first feature point information;
• and forming the corresponding target region in the image to be detected based on the feature points;
• wherein the target region is a closed polygon formed by connecting the feature points.