Composite Image Subject Detection via Pre-Compositing Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques fail to accurately detect and recognize subjects in composite images generated from multiple material images, as subjects from individual images may overlap, making it difficult to correctly perform detection and recognition.
Innovation Solution
An image processing apparatus that obtains subject information from individual images, generates a composite image by compositing these images, and records the subject information associated with the composite image, including a detection unit to detect subjects within the composite image and generate corresponding subject information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a composite image is generated by compositing multiple material images, then the visual composition is enhanced, but the accuracy of subject detection and recognition deteriorates due to overlapping subjects
Solution Approach 1:
The patent applies preliminary action by detecting and recording subject information from each material image before the compositing process occurs. The detection unit identifies subjects in individual material images and generates subject information including position, type, and other attributes. This pre-detection ensures that subject data is captured before overlapping in the composite image makes detection difficult, thereby resolving the contradiction between composite image generation and detection accuracy.
Solution Approach 2:
The patent segments the subject detection process into two independent stages: first detecting subjects in individual material images separately, then compositing the images. This segmentation allows each material image to be analyzed independently for subject information, avoiding the confusion caused by overlapping subjects in the final composite image, thus maintaining high detection accuracy while enabling composite image generation.
2Device complexity
If subject detection is performed only on the composite image, then the processing flow is simplified, but subject information from individual material images is lost
Solution Approach 1:
The patent performs subject detection as a preliminary action on each material image before compositing, rather than only on the final composite image. The detection unit processes each material image independently to extract subject information, which is then recorded and associated with the composite image. This approach prevents information loss while maintaining relatively simple processing through automated workflows.
Solution Approach 2:
The patent introduces an intermediary mechanism in the form of subject information records that bridge the material images and the composite image. The recording unit stores subject information from each material image, and this intermediate data structure preserves subject details even after compositing. This intermediary layer ensures no subject information is lost while keeping the overall system manageable.
3Measurement precision
If subject information from multiple material images is recorded in association with the composite image, then subject recognition accuracy is improved, but the data storage requirements increase
Solution Approach 1:
The patent uses copying by creating compact subject information records that contain essential attributes (position, type, identification) of detected subjects. Instead of storing complete material images or redundant data, the system generates concise copies of subject data that can be efficiently stored and associated with the composite image, thereby improving recognition accuracy without proportionally increasing storage requirements.
Solution Approach 2:
The patent applies discarding and recovering by selectively retaining only the essential subject information attributes needed for recognition (position, type, ID) while discarding redundant or unnecessary data. The recording unit stores this filtered subject information in an optimized format, recovering only the critical data needed for accurate subject recognition in the composite image context, thus balancing accuracy with storage efficiency.
Data Source
AI summary
There is provided an image processing apparatus. An obtainment unit obtains a first image, first subject information indicating a first subject detected from the first image, a second image, and second subject information indicating a second subject detected from the second image. A composition unit generates a composite image by compositing the first image and the second image. A recording unit records the first subject information and the second subject information in association with the composite image.