Human face snapshot method and system
A face image and face detection technology, which is applied in the field of computer vision, can solve the problems of the small available range of the face capture system and the inability to realize long-distance and large-scale scene face capture, etc.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0086] figure 1 It shows a schematic flow diagram of the implementation of the face capture method in Embodiment 1 of the present application. As shown in the figure, the face capture method may include the following steps:
[0087] Step 101, using the wide-angle camera to obtain wide-angle video data in the monitoring scene;
[0088] Step 102, detecting the area where the human body is located according to the video data;
[0089] Step 103, using a telephoto camera to obtain video data of the area where the human body is located;
[0090] Step 104: Perform face detection on the video data acquired by the telephoto camera to obtain a face image.
[0091] In a specific implementation, the wide-angle camera may have a camera lens (wide-angle lens) with a very wide viewing angle, which can accommodate more scenes within a limited distance. The parameter to measure the camera’s wide-angle is usually the minimum focal length. Generally, the focal length of a wide-angle lens is b...
Embodiment 2
[0142] In order to achieve ultra-long-distance face capture, this application can use the wide-angle camera of the linked camera group to conduct a preliminary screening of the area of interest, and then call the telephoto camera to shoot high-definition ultra-long-distance video for the area of interest for high-resolution Accurate face capture.
[0143] figure 2 A schematic diagram of the face capture process in Embodiment 2 of the present application is shown. As shown in the figure, the face capture process may include the following steps:
[0144] Step 201, performing human body detection on the field of view of the wide-angle camera;
[0145] Get pictures from the wide-angle camera, use the deep neural network for human detection, and conduct a preliminary screening of the areas where the target (face) may appear (the place where the human body appears) at an extensive angle.
[0146] Step 202, judging whether the region of interest is found;
[0147] If found, e...
Embodiment 3
[0173] The following embodiments of the present application describe the deep neural network structure involved in human body / face detection.
[0174] image 3 A schematic diagram of the human body / or face detection process in Embodiment 3 of the present application is shown. As shown in the figure, the human body or face detection process may include the following steps:
[0175] Step 301, perform preprocessing on the human body or human face image.
[0176] Preprocess the picture (the input picture for human body detection can be taken by a wide-angle camera, and the input picture for face detection can be taken by a telephoto camera): scale and adjust it to a reasonable size (assumed to be 300*500), and adjust it to a uniform color Space (such as BGR), and subtracting the pixel value of the average human body / face, etc.
[0177] Step 302, perform multiple convolution calculations on the image to obtain a feature image.
[0178] The input picture is input to multiple conv...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More