Horizontal-to-Vertical Image Conversion with Subunit AI Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image reproduction technologies result in reduced visual satisfaction due to excessive use of letter boxes or pillar boxes when horizontally long images are reproduced on vertically long screen devices, leading to inefficient use of screen space and degraded image quality.
Innovation Solution
An electronic device and system that utilize AI models to analyze and separate images into subunits, calculate optimal reproduction areas, and apply AI models to enhance and convert images from a horizontally long to a vertically long format, ensuring full-screen reproduction with improved image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If conventional screen ratio conversion technology is used to reproduce horizontal images on vertical screens, then the image ratio is maintained, but excessive letter boxes or pillar boxes are generated reducing visual satisfaction
Solution Approach 1:
The patent segments the horizontal image into multiple vertical sub-images, then processes and reassembles them to create a vertical output image. This segmentation allows the system to transform the image orientation while maintaining content integrity and eliminating black borders.
Solution Approach 2:
The patent applies AI-based super-resolution technology to generate additional vertical pixels, effectively adding a new dimension to the image transformation process. This enables the conversion from horizontal to vertical format while enhancing image quality and filling the entire screen area.
2Area of stationary object
If the image is reduced to fit the vertical screen ratio, then the screen space is utilized, but the image quality is degraded
Solution Approach 1:
The patent uses AI super-resolution to generate new vertical pixels, effectively expanding the image in the vertical dimension without compromising quality. This allows full-screen utilization while maintaining or even enhancing image fidelity through intelligent pixel generation.
Solution Approach 2:
The patent changes the resolution parameters vertically by applying super-resolution algorithms that generate high-quality pixels. This transforms the image resolution from a simple reduction operation to an intelligent parameter transformation that maintains quality while adapting to the vertical screen format.
3Manufacturing precision
If AI models are applied to each subunit for image transformation, then image quality is enhanced, but processing complexity increases
Solution Approach 1:
The patent divides the image into subunits and applies AI models selectively to each segment. This segmentation strategy reduces the overall processing complexity by breaking down a large complex transformation into smaller, more manageable tasks that can be processed independently and efficiently.
Solution Approach 2:
The patent applies different AI processing strategies to different regions of the image based on local characteristics. By identifying important regions and applying appropriate super-resolution techniques selectively, the system enhances image quality where needed while reducing unnecessary processing in other areas, thereby managing complexity.
Data Source
AI summary
Proposed is an electronic device, a system, and a method for intelligent horizontal-vertical image conversion. The device may transmit a bitstream containing information on an image having a first image ratio that is longer horizontally than vertically to a terminal to enlarge and reproduce the image when the terminal has a screen ratio state that is longer vertically than horizontally. The device may include an analysis controller for analyzing contents of a corresponding frame image to calculate a corresponding reproduction area. The device may also include a selection controller for separating the image into a plurality of subunits, and selecting an optimal artificial intelligence (AI) model applied for each subunit according to the contents of the image within the corresponding subunit from among a plurality of previously trained AI models. The device may further include a generation controller for generating the bitstream, the reproduction area, and the optimal AI model.