Video Encoding Region of Interest Correction via Motion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving image encoding techniques, such as those using H.264 and H.265, often result in low image quality for regions of interest when encoded at low bit rates, as they rely on pre-defined region settings and motion vectors, leading to potential misclassification and reduced accuracy in determining important regions.
Innovation Solution
A moving image encoding apparatus that includes a detection unit for motion information, a determination unit to correct region of interest based on motion vectors, and a control unit to set quantized values for improved image quality, ensuring accurate identification and high-quality encoding of regions of interest while reducing bit rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If encoding is performed at a low bit rate to suppress the size of encoded data, then the encoding amount is reduced, but the image quality deteriorates and important regions cannot be recognized
Solution Approach 1:
The patent applies local quality by differentiating encoding precision between regions of interest and non-interest areas. Regions of interest (such as moving objects or persons) are encoded with high image quality using lower quantized values, while non-interest regions are encoded with suppressed quality using higher quantized values. This allows the system to maintain acceptable overall image quality while significantly reducing the total encoded data size.
2Measurement precision
If a region is determined in advance as a region of interest using coordinate information data, then the region can be distinguished with high precision, but if the region of interest is included in the region determined as a region of non-interest, the region of interest cannot be recognized
Solution Approach 1:
The patent implements feedback by using motion detection results to verify and correct the initial region of interest determination. After initially identifying regions of interest based on coordinate information data, the system performs motion detection and uses the motion vectors to confirm whether the identified regions actually contain moving objects. This feedback mechanism ensures that regions of interest are accurately recognized and prevents misclassification.
Solution Approach 2:
The patent applies preliminary action by performing motion detection and generating motion vectors for the entire image before finalizing the region of interest determination. This preliminary motion analysis is then used to verify and adjust the initially determined regions of interest, ensuring accurate identification before the encoding process begins.
3Productivity
If the entirety of a frame is not encoded to suppress encoding amounts, then the bit rate is reduced, but information loss increases and original functions are impaired
Solution Approach 1:
The patent applies local quality by encoding different regions of the frame with different quality levels. Regions of interest are encoded with high quality to preserve important information, while non-interest regions are encoded with lower quality or lower resolution to reduce data量. This selective encoding approach maintains encoding efficiency while minimizing information loss in critical areas.
Data Source
AI summary
A moving image encoding apparatus comprises a detection unit configured to detect motion information in units of blocks from a moving image; a determination unit configured to determine a region of interest in the moving image based on a first region determined through processing for detecting an object from an image, and the motion information; a control unit configured to perform control such that a quantized value of a block determined as being the region of interest is set to a value lower than a quantized value of a block determined as not being the region of interest; and an encoding unit configured to perform compression encoding on the moving image based on the quantized value set by the control unit.