Image Capture Sub-Image Stitching for High Resolution Output
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile computing devices with high-resolution image sensors are limited by hardware constraints such as memory capacity and latency, preventing them from outputting images at their full capture resolution, resulting in lower resolution images that cannot be used for large prints or desired field of view.
Innovation Solution
The method involves capturing multiple sub-images at lower resolutions than the desired output resolution, which are then stitched together to form an output image at the desired resolution, circumventing hardware limitations and allowing for higher resolution images than the device's output capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the image sensor captures at full resolution, then image quality and field of view are improved, but hardware limitations (memory capacity, latency) prevent output at that resolution
Solution Approach 1:
The patent divides the full-resolution scene into multiple lower-resolution sub-images that can be captured and processed within hardware limitations. These sub-images are then stitched together to reconstruct the full-resolution output, effectively segmenting the imaging process to bypass memory and latency constraints while maintaining final image quality.
2Measurement precision
If multiple sub-images are captured and stitched, then output resolution exceeds device capabilities, but processing complexity increases
Solution Approach 1:
The patent introduces an intermediary processing system that automatically handles the capture, stitching, and reconstruction of multiple sub-images into a single high-resolution output. This intermediary layer manages the complexity of processing multiple images, providing a seamless solution that achieves super-resolution without requiring direct user involvement in the complex stitching process.
3Measurement precision
If hardware is upgraded to support full resolution output, then image quality improves, but device size and cost increase
Solution Approach 1:
The patent transitions from a single-dimension approach (capturing one full-resolution image) to a multi-dimensional approach (capturing multiple lower-resolution sub-images from different positions or angles and combining them). This dimensional change in the imaging process allows the system to achieve full-resolution output without requiring proportional increases in hardware capacity, thereby maintaining compact device size.
Data Source
AI summary
Methods and apparatuses are provided for capturing an image. A method may include determining a scene to be captured at a first resolution. The scene to be captured may be within a field of view of an image sensor. The method may further include causing capture by the image sensor of a plurality of sub-images at a second resolution. The second resolution may be less than the first resolution. Each of the plurality of sub-images may depict a portion of the scene to be captured. The method may additionally include using the plurality of sub-images to generate an output image depicting the scene to be captured at the first resolution. Corresponding apparatuses are also provided.