ML Image Aspect Ratio Adaptation With GAN Extension and ROI Cropping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Images with mismatched aspect ratios between the image and display area result in undesirable white space, distortion, or loss of significant visual features when presented, such as in advertisements.
Innovation Solution
A generative adversarial network (GAN) is used to extend images with additional features, and a discriminator is employed to ensure consistency and realism, followed by automatic cropping to fit the display aspect ratio using machine learning techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If the image aspect ratio is changed to fit the display area, then the image fills the display area without white space, but the image becomes distorted or loses visual features
Solution Approach 1:
The system performs preliminary actions by generating multiple candidate aspect ratios and pre-processing the image through various transformations (flipping, rotating, cropping) before the final aspect ratio selection. This allows the system to explore multiple possibilities and select the best matching aspect ratio that preserves visual features while fitting the display area.
Solution Approach 2:
The system uses feedback mechanisms by evaluating each candidate aspect ratio against the original image using machine learning models that assess visual feature preservation. The system iteratively refines the aspect ratio selection based on feedback from image analysis, ensuring the chosen aspect ratio maintains the highest possible image fidelity while filling the display area.
2Adaptability or versatility
If the image is cropped to fit the display area, then the aspect ratio matches, but significant visual features such as text or objects are lost
Solution Approach 1:
The system dynamically adjusts the cropping strategy by identifying regions of interest (ROIs) containing significant visual features and adapting the crop boundaries to preserve these features. The cropping process is not static but dynamically optimized based on the detected importance of different image regions, allowing the system to maintain aspect ratio adaptability while minimizing information loss.
Solution Approach 2:
The system applies local quality preservation by differentiating between important and less important regions of the image. It identifies regions containing significant visual features (text, objects, faces) and ensures these regions are preserved in the cropped output, while less important areas can be adjusted to achieve the target aspect ratio. This selective preservation maintains visual feature integrity during adaptation.
3Manufacturing precision
If machine learning techniques are used to identify regions of interest, then visual feature preservation is improved, but processing time and computational resources increase
Solution Approach 1:
The system segments the image processing task into distinct stages: initial quick analysis to identify potential regions of interest, detailed analysis of identified regions to confirm their importance, and final aspect ratio determination. This segmentation allows the computationally intensive machine learning operations to be focused only on relevant image regions rather than processing the entire image uniformly, reducing overall processing time while maintaining high accuracy in region identification.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
To adjust an aspect ratio of an image to match the aspect ratio of a display area for presenting the image, a computing device receives an image having a first aspect ratio, and obtains a second aspect ratio for a display area of a display in which to present the image, where the second aspect ratio is different from the first aspect ratio. The computing device extends the image to include one or more additional features which were not included in the image. Additionally, the computing device automatically crops the extended image around an identified region of interest by selecting a portion of the extended image that has an aspect ratio which matches the second aspect ratio of the display area, and provides the cropped image for presentation within the display area of the display.