Image Texture Conversion Using Object Masks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing methods for texture conversion and synthesis lack control over conversion contents and positions, leading to undesirable results such as object distortion and content duplication, resulting in poor composite images.
Innovation Solution
An image processing method that uses object masks to extract and process images with a Visual Geometry Group Network (VGG) series model, calculating object and background losses to optimize texture conversion, allowing for precise control over area selection and reducing redundant content, thereby improving image accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If texture conversion and synthesis algorithms are used to extract content information and modify images, then image modification capability is improved, but control over conversion contents and positions deteriorates, leading to object distortion and content duplication
Solution Approach 1:
The patent segments the image processing task by introducing object masks that divide the image into distinct object regions and background regions. This segmentation allows independent control over different areas, enabling precise control of texture conversion positions and contents while maintaining the overall image modification capability. The mask-guided approach separates the conversion process into region-specific operations, preventing unwanted distortion and duplication.
Solution Approach 2:
The patent applies local quality by allowing different processing strategies for different regions of the image. Object regions identified by masks receive targeted texture conversion while background regions are preserved or processed differently. This localized control ensures that conversion contents and positions are precisely controlled in critical areas, eliminating object distortion and content duplication while maintaining versatility in image modification.
2Manufacturing precision
If object masks are extracted and used to guide texture conversion, then control precision over conversion positions is improved, but processing complexity increases due to mask extraction and region division
Solution Approach 1:
The patent applies preliminary action by extracting object masks before the texture conversion process. This pre-processing step identifies and segments object regions in advance, allowing the main conversion algorithm to operate with clear guidance on where conversions should occur. This preliminary segmentation reduces the complexity of the conversion process itself, as the algorithm only needs to apply textures to predefined regions rather than determining positions dynamically.
Solution Approach 2:
The patent introduces object masks as an intermediary element between the input image and the texture conversion process. These masks serve as a mediator that guides the conversion algorithm, simplifying the interaction between the processing system and the image data. The masks act as a control layer that reduces computational complexity by pre-defining conversion regions, allowing the main algorithm to focus on texture application rather than position determination.
3Adaptability or versatility
If the entire image is processed for texture conversion, then comprehensive modification is achieved, but processing time increases and efficiency decreases
Solution Approach 1:
The patent segments the image processing task by dividing the image into object regions and background regions using masks. This segmentation enables selective processing where texture conversion is applied only to relevant object regions rather than the entire image. The comprehensive modification is maintained through proper mask design that covers all necessary areas, while processing efficiency is improved by avoiding unnecessary computations in background regions.
Solution Approach 2:
The patent applies partial action by processing only the necessary portions of the image identified by object masks. Instead of applying texture conversion to the entire image, the algorithm processes only the masked object regions, reducing overall processing time and computational load. This selective approach maintains comprehensive modification coverage for all important image elements while significantly improving processing efficiency by eliminating redundant operations.
Data Source
AI summary
An image processing method and an electronic device are provided, the method extracts a first object mask of a texture image and a second object mask of a to-be-optimized image. An image recognition model is used to obtain a first content matrix, a first texture matrix, a second content matrix, a second texture matrix, a first mask matrix, and a second mask matrix. A total loss of the to-be-optimized image is determined, and the total loss is minimized by adjusting a value of each pixel of the to-be-optimized image, thereby an optimized image is obtained. By utilizing the image processing method, quality of final image is improved.