Video Conference Speaker Framing Using Voice-Guided Image Areas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional auto framing systems struggle to correctly focus on the target object in panoramic images due to multiple faces, leading to ineffective framing in video conferences.
Innovation Solution
A video conference system that utilizes audio and image processing to detect the direction of voice signals, perform image recognition on specific image areas, and display an emphasized frame highlighting the object of interest.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional auto framing system is applied to panoramic image, then the system can process the image, but it cannot correctly focus on the target object due to too many faces
Solution Approach 1:
The patent divides the panoramic image into multiple image areas (first image area, second image area, third image area, fourth image area) based on directional information from voice signals. This segmentation allows the system to focus on specific regions where the target object is located, rather than processing the entire panoramic image at once, thereby improving target object focus accuracy while maintaining ease of operation.
Solution Approach 2:
The patent introduces voice signal directional information as an intermediary to guide the image recognition process. The control device uses the direction of the voice signal to determine which image area contains the target object, enabling accurate focusing even in panoramic images with multiple faces. This intermediary mechanism resolves the contradiction by providing a reliable cue for target identification.
2Productivity
If the system processes the entire panoramic image, then all image areas are analyzed, but the target object cannot be correctly identified among multiple faces
Solution Approach 1:
The patent segments the panoramic image into multiple directional image areas and uses voice signal direction to select the relevant area for analysis. This approach maintains productivity by processing only the necessary portion of the image rather than the entire panorama, while improving target object identification accuracy by eliminating distractions from other faces in different directions.
Solution Approach 2:
The patent applies local quality by focusing the image recognition process on the specific image area where the target object is located, rather than uniformly processing the entire image. The control device adjusts the processing focus based on voice signal direction, ensuring high identification accuracy in the relevant local area while maintaining overall processing efficiency.
Data Source
AI summary
The present disclosure provides video conference system and conference frame display method. The conference frame display method includes: detecting a direction of at least one voice signal according to at least one audio signal; performing an image recognition on at least one of a plurality of image areas in a rectangle image according to the direction of the at least one voice signal; and in a condition that at least one object of interest is recognized from the at least one of the plurality of image areas, by a display device, displaying an emphasized frame visually highlighting the at least one object of interest.