Target Object Image Splicing With Template Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing applications do not adequately meet users' personalized requirements for interactive functions, particularly in image splicing.
Innovation Solution
An image splicing method and apparatus that acquires a to-be-spliced image and multiple template images, calculates the matching degree between target objects, selects a template image based on a preset condition, and splices corresponding image portions to create a personalized target spliced image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image splicing methods are used, then the splicing process is simple, but the matching accuracy between target objects is low
Solution Approach 1:
The patent segments the image processing task into distinct modules: target object detection, feature extraction, matching degree calculation, and splicing execution. This segmentation allows each module to be optimized independently, improving matching accuracy through sophisticated algorithms while managing system complexity through modular architecture.
Solution Approach 2:
The patent performs preliminary actions by pre-processing images to extract target objects and their features before the actual splicing operation. The matching degree is calculated in advance, and the best template is selected beforehand, which improves the accuracy of the final splicing result while organizing the complexity into manageable preprocessing steps.
2Manufacturing precision
If multiple template images are processed to improve matching accuracy, then the splicing quality improves, but the processing time increases
Solution Approach 1:
The patent calculates matching degrees for multiple template images (excessive action) but only selects the top matching template for splicing (partial action). This approach ensures high splicing quality by evaluating multiple options while avoiding the time cost of processing all templates equally, as only the best match is used in the final result.
Solution Approach 2:
The patent changes parameters by adjusting the preset threshold for matching degree. By setting an appropriate threshold, the system can quickly filter out templates that don't meet the quality standard, reducing the number of templates that require detailed processing while maintaining high splicing quality for the selected template.
3Adaptability or versatility
If automatic matching is implemented to satisfy personalized requirements, then user experience improves, but the computational complexity increases
Solution Approach 1:
The patent implements self-service by enabling the system to automatically detect target objects, calculate matching degrees, and select the best template without user intervention. This automatic matching process provides personalized splicing results based on the user's input image, improving user experience while managing computational complexity through efficient algorithms and preset thresholds.
Solution Approach 2:
The patent replaces manual image matching (mechanical system) with automated computer vision algorithms. The system uses feature extraction and matching degree calculation to automatically identify the best template, substituting the need for manual selection with computational processes that provide personalized results efficiently.
Data Source
AI summary
Provided are an image splicing method, apparatus and device, and a medium. The image splicing method includes acquiring a to-be-spliced image and multiple template images, where the to-be-spliced image and each template image each include a target object; calculating a matching degree between the target object in the to-be-spliced image and the target object in the each template image; determining a template image whose matching degree satisfies a preset matching condition as a target template image; and splicing a target image portion in the to-be-spliced image to a target image region in the target template image to obtain a target spliced image, where the target image portion and the target image region correspond to the same portion of the target object.