Image Processing With AI Intermediate Images for Live-View Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies fail to adequately reduce the sense of incongruity when displaying reconstructed images, particularly due to significant differences between live view and recorded images caused by timing deviations during shooting and recording.
Innovation Solution
An image processing apparatus that utilizes a trained learning model to generate images with adjusted fluctuation elements, displaying the original image if divergence exceeds a threshold, and generating a new image with reduced fluctuation when necessary, to minimize the perceived difference between the live view and reconstructed images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a reconstructed image is generated using AI technologies, then the user's intention is reflected in the image, but the divergence between the recorded image and the reconstructed image increases
Solution Approach 1:
The patent introduces an intermediate image as a mediator between the recorded image and the reconstructed image. This intermediate image serves as a transition step that reduces the sense of incongruity while still reflecting user intentions. The display control unit displays the intermediate image when the divergence between recorded and reconstructed images exceeds a threshold, thereby bridging the gap between the two extreme states.
Solution Approach 2:
The patent changes the parameter of image fluctuation by generating an intermediate image with adjusted fluctuation characteristics. The learning model is trained to generate images with controlled fluctuation parameters, allowing the system to produce an intermediate representation that maintains user intention while reducing visual divergence from the original recorded image.
2Ease of operation
If smoothing filter processing is applied to reduce timing deviation effects, then the sense of incongruity is reduced, but the image quality and detail are degraded
Solution Approach 1:
The patent replaces traditional mechanical smoothing filter processing with an AI-based learning model approach. Instead of applying conventional signal processing techniques that degrade image quality, the system uses a trained learning model to generate an intermediate image that reduces timing deviation effects while preserving image quality and details through intelligent synthesis rather than mechanical filtering.
3Ease of operation
If multiple manipulated images are displayed sequentially, then the transition between live view and recorded image is smoother, but the display time and user waiting period increase
Solution Approach 1:
The patent applies partial action by generating and displaying only one intermediate image rather than multiple manipulated images sequentially. This intermediate image provides sufficient transition effect to reduce the sense of incongruity while minimizing the additional display time required, avoiding the excessive time consumption of displaying multiple intermediate steps.
Data Source
AI summary
An image processing apparatus comprises: an image acquisition unit that acquires a first image; a degree-of-fluctuation acquisition unit that acquires a degree of fluctuation of a fluctuation element among elements constituting the first image; a generation unit that uses the first image to generate a second image in which the degree of fluctuation of the fluctuation element is different from the first image with use of a trained learning model; and a display control unit that controls to display an image. If the first image and the second image diverge by a predetermined divergence or more, the generation unit further uses the first image to generate a third image in which the degree of fluctuation of the fluctuation element is less than the second image, and the display control unit controls to display the first image and to thereafter display the third image.