Video Object Identification Using Gyro-Based Rotation Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electronic devices struggle to accurately identify objects in videos when the device is rotated, as the rotation introduces errors in object recognition and distance estimation.
Innovation Solution
The electronic device employs a combination of a gyro sensor and a designated estimation model to obtain rotation data, which is then used to calculate a rotation value for the video information. This rotation value allows the device to correctly identify objects in the video by rotating the video data accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the electronic device uses a camera to obtain video information, then the device can capture visual data, but when the device is rotated the object identification accuracy deteriorates
Solution Approach 1:
The patent introduces rotation data as an intermediary element that mediates between the camera's video capture and the object identification process. The rotation data, obtained from gyro sensors and estimation models, serves as a bridge that allows the system to compensate for device rotation and maintain accurate object identification regardless of device orientation.
Solution Approach 2:
The patent changes the parameter of video information by applying rotation compensation based on obtained rotation data. The processor modifies the video parameters (rotation angle, orientation) to counteract the device's physical rotation, thereby maintaining consistent object identification accuracy across different device orientations.
2Measurement precision
If the electronic device combines gyro sensor and estimation model to obtain rotation data, then the rotation value accuracy improves, but the device complexity increases
Solution Approach 1:
The patent merges two different approaches to obtaining rotation data: direct measurement from gyro sensors and indirect estimation from video information using an estimation model. By combining these complementary methods, the system achieves higher rotation value accuracy than either method alone, while distributing the computational load across different components.
Solution Approach 2:
The patent makes the rotation data acquisition system multi-functional by using the same infrastructure (processor, memory) to both capture video information and process gyro sensor data. The estimation model serves multiple purposes: it processes video data and simultaneously provides rotation estimation, reducing the need for dedicated hardware components.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables the electronic device to smoothly recognize objects in videos even when the device is rotated, improving the accuracy of object identification and distance estimation.
Implementation Method 1
obtaining, first rotation data on the electronic device by using a gyro sensor of the electronic device
Data Source
AI summary
According to an embodiment, an electronic device includes at least one camera, a gyro sensor, memory, and at least one processor operably connected with the at least one camera, the gyro sensor, and the memory, and the processor is configured to obtain video information by using the at least one camera, based on obtaining the video information, obtain, first rotation data on the electronic device by using the gyro sensor, and second rotation data by using a designated estimation model, based on the first rotation data and the second rotation data, obtain a rotation value for the video information, and based on the obtained rotation value, identify at least one object regarding the video information.