Image Processing Apparatus for Distortion-Corrected Encoding Area Setting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies struggle to set areas associated with encoding functions in distortion-corrected images from omnidirectional cameras, such as fish-eye images, which are essential for improving image quality in specific regions, as these areas become distorted during geometric conversion, making it difficult to apply high-quality image settings in a manner compliant with encoding standards.
Innovation Solution
An image processing apparatus and method that allows for the separate setting of areas influencing encoding processing in both the original fish-eye image and the distortion-corrected cut image, using user interfaces to define high-quality image areas within these images, enabling encoding based on user-defined settings that align with encoding units like macroblocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If geometric conversion processing is performed to correct distortion in fish-eye images, then image quality is improved, but the area associated with encoding functions cannot be properly set due to shape distortion
Solution Approach 1:
The patent introduces a coordinate transformation mechanism that maps the distorted area in the corrected image back to the original fish-eye image coordinate system. This dimensional mapping allows the encoding area setting to be performed in the undistorted original coordinates while applying to the corrected image, resolving the contradiction between distortion correction and area setting capability.
2Manufacturing precision
If quantization step width is reduced to increase image quality in specific areas, then image quality improves, but code amount increases
Solution Approach 1:
The patent enables different quantization step widths to be applied to different areas of the image. By allowing users to specify high-quality areas in the original fish-eye image coordinates, the system applies smaller quantization steps only to those specific regions while using larger steps in other areas, thus improving local image quality without proportionally increasing the overall code amount.
3Area of stationary object
If the high-quality image area is expanded in the cut image, then more area benefits from high quality, but the area becomes distorted and cannot be properly encoded
Solution Approach 1:
The patent performs coordinate transformation and area mapping in advance, before the encoding process. By calculating the corresponding area in the original fish-eye image that maps to the desired area in the corrected image, the system pre-determines the encoding parameters for each macroblock, ensuring proper encoding compliance even for expanded high-quality areas.
Data Source
AI summary
An image processing apparatus obtains a first image, and cuts out at least a portion of the first image to generate a second image that is distortion-corrected. The image processing apparatus provides a first user interface for setting, in the first image, an area associated with a function that influences encoding processing, and a second user interface for setting, in the second image, an area associated with the function that influences encoding processing. The image processing apparatus encodes the first image based on the area set using the first user interface, and encodes the second image using the area set using the second user interface.