Image Layout Apparatus Face Size Template Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for laying out images on a single background struggle with images containing small faces, leading to enlargement issues that distort the original mood and background prominence, especially when faces are extremely small or absent, and varying image insertion regions complicate discernment of individuals in the image.
Innovation Solution
An image layout apparatus and method that selects images based on face presence, using templates with varying image insertion regions, where images with faces are inserted into larger regions and those without faces into smaller regions, ensuring appropriate sizing and maintaining the intended image composition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If faces are normalized to predetermined size, then uniformity of facial images is improved, but the mood of the image is lost and background is excluded
Solution Approach 1:
The patent applies local quality by differentiating the treatment of images based on face size. Images with large faces undergo normalization to achieve uniformity, while images with small faces are preserved in their original state to maintain the background and mood. This selective approach allows different parts of the image collection to have different processing qualities appropriate to their content.
Solution Approach 2:
The patent changes the parameter of image size normalization based on the detected face size. Instead of applying a fixed normalization rule to all images, the system adjusts the normalization parameter dynamically: images with faces below a certain size threshold are exempt from normalization, while others are normalized to a predetermined size. This parameter adaptation resolves the contradiction between uniformity and information preservation.
2Length of moving object
If images with small faces are enlarged, then face size is improved, but the background and original composition are distorted
Solution Approach 1:
The patent applies preliminary anti-action by detecting face size before the layout process and preemptively deciding not to enlarge images with small faces. This prevents the harmful effect of distortion before it occurs. By identifying images with faces below the threshold in advance, the system avoids the enlargement operation that would otherwise distort the background and composition, thus counteracting the potential harm before it manifests.
3Quantity of substance
If images are inserted into small regions, then layout density is improved, but discernment of individuals becomes difficult
Solution Approach 1:
The patent changes the layout parameter by adjusting image size based on the detected face size. Images with small faces are inserted at their original size rather than being scaled down to fit small regions, ensuring that faces remain discernible. This dynamic parameter adjustment allows the system to maintain both high layout density where appropriate and sufficient face visibility where needed, resolving the contradiction between quantity and discernment.
4Manufacturing precision
If face normalization is applied to all images, then uniformity is improved, but images without faces are improperly processed
Solution Approach 1:
The patent applies local quality by making the normalization process conditional on the presence and size of faces. Images are examined individually, and only those containing faces of sufficient size undergo normalization. This selective application ensures that uniformity is achieved where applicable while avoiding incorrect processing of images without faces, thus maintaining both uniformity and processing accuracy.
Solution Approach 2:
The patent uses feedback from face detection results to control the normalization process. The face detection module provides feedback about the presence and size of faces in each image, which then determines whether normalization should be applied. This feedback mechanism ensures that images are processed appropriately based on their actual content, improving both uniformity and processing reliability.
Data Source
AI summary
Images are laid out appropriately according to the sizes of faces of people pictured therein. A user selects a plurality of images to be laid out on a single background image. A layout determining section selects a template having the same number of image insertion regions as the number of images selected by the user, from among a plurality of templates, each having one or more image insertion regions of various sizes. Whether faces are included in the images is judged. Image layout is determined such that images that include faces are inserted in image insertion regions of the template having greater sizes as the size of a representative face is smaller; and images that do not include faces are inserted into image insertion regions of the template having smaller sizes.