An automatic focusing method of face area in the state of large magnification and shallow depth of field

Abstract

The present invention provides an automatic focusing method for a human face area in the state of large magnification and shallow depth of field. In the scene of large magnification and shallow depth of field, face detection, area optimization, and automatic focusing can be performed in camera products with insufficient computing power to find the speaker. The location of the Focus motor is the clearest in the area. The present invention comprises the following steps: Step 1, training a face detector based on the YOLO V3 neural network; Step 2, using the face detector to calculate the set of human face regions in the image: Step 3, optimizing the human face region; Step 4, optimizing the human face The face area is matched with the area divided by the chip, and the focus area of ​​interest is locked; step five, calculate the area of ​​interest: step six, automatic focus.

Description

technical field [0001] The invention relates to an automatic focusing method for a human face area in a state of large magnification and shallow depth of field, and is applied in the field of cameras for education recording and broadcasting and video conferencing. Background technique [0002] In recent years, in applications such as educational recording, video conferencing, etc., 10x, 12x, or 20x zoom cameras are usually used to play videos with people as the theme in real time in an indoor scene of 5-10 meters. The key point is to ensure that the facial areas of the characters are absolutely clear. In this context, it is an effective way to solve this problem by proposing to use the face area as the key area for automatic focusing. [0003] Considering such a usage scenario, such as figure 1 As shown, in this three-dimensional scene, the sequence from far to near is the background (blackboard, background wall, etc.), the subject (speaker), and the camera (10 times, 12 t...

AI Technical Summary

Problems solved by technology

[0004] (1) The depth of field will become shallow when the zoom lens is pulled to a large magnification, and it is impossible to ensure that the background and characters are clear at the same time

[0005] (2) The strategy of reducing noise and increasing the aperture will lead to a shallower depth of field, which cannot ensure that the background and characters are clear at the same time

[0006] (3) The background details are too rich, which is mistaken for the subject of the image in the autofocus algorithm and a false peak is calculated, resulting in a clear background and blurred characters

[0007] (4) The details of the characters are too single, and the weight in the auto-focus algorithm is too small, which is ignored by the algorithm when calculating FV (Focus Value), resulting in clear distances from other objects and blurred characters

However, in the face of well-designed, compact and cost-effective camera products, part of the computing power is often consumed in terms of imaging requirements, and then embedding face detection and area designation algorithms into the system will inevitably cause a shortage of computing power, making it difficult to achieve real-time processing.

Images