Computational Photography Architecture with Reusable Algorithm Modules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The current computational photography systems face challenges in algorithm module development due to independent development by different developers, leading to long development times and low efficiency, lack of functional sharing, inflexible pipelines, limited high-level APIs, and inefficient inter-process communication.
Innovation Solution
A data processing method and system that includes an auxiliary unit and image processing unit, with separate independent running modules, enabling plugin management and reusability of algorithm modules, and utilizing inter-process communication for efficient image processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If algorithm modules are independently developed by respective developers, then each application can have customized algorithm functions, but algorithm modules cannot be reused between applications, resulting in long development time and low development efficiency
Solution Approach 1:
The patent creates a universal algorithm module management mechanism where algorithm modules can serve multiple applications simultaneously. The system provides common algorithm functions (such as image processing, object detection, etc.) that can be reused across different applications, eliminating the need for each application to independently develop the same algorithm modules. This resolves the contradiction by making algorithm modules universal while maintaining the ability to customize through selective composition.
Solution Approach 2:
The patent segments the algorithm module system into independent, reusable components that can be individually developed, tested, and deployed. Algorithm modules are divided into functional units that can be independently managed and combined differentially for different applications. This segmentation enables both customization (by selecting different segments) and reusability (by sharing segments across applications).
2Adaptability or versatility
If algorithm modules are independently developed, then each application has its own algorithm implementation, but functional sharing between applications is impossible
Solution Approach 1:
The patent merges the algorithm module management across multiple applications into a unified system. Common algorithm functions are consolidated into shared algorithm modules that can be invoked by multiple applications. This merging enables functional sharing while maintaining application-specific customization through selective composition of shared modules, thus resolving the contradiction between application-specific implementation and functional sharing.
3Reliability
If computational photography algorithms are integrated in the Camera HAL or application layer, then algorithm functions can be implemented, but the development process is complex and algorithm modules cannot be easily reused
Solution Approach 1:
The patent introduces an intermediary algorithm module management layer between the Camera HAL and application layers. This intermediary layer provides standardized interfaces and mechanisms for algorithm module loading, execution, and management. By inserting this intermediary layer, the system achieves reliable algorithm function implementation while simplifying the development process through standardized module interfaces and reuse mechanisms, thus resolving the contradiction between reliability and complexity.
4Adaptability or versatility
If algorithm modules are independently developed by different developers, then diverse algorithm functions can be created, but development time is extended and efficiency is reduced
Solution Approach 1:
The patent implements preliminary action by pre-defining standardized algorithm module interfaces, management mechanisms, and common algorithm functions. Developers can reuse these pre-prepared components and templates, significantly reducing development time for new algorithm functions. The standardized framework allows diverse algorithm functions to be created quickly by composing existing modules rather than developing from scratch, thus resolving the contradiction between diversity and development time.
Data Source
AI summary
Provided are a data processing method, a data processing system, an electronic device, and a storage medium. The data processing method is applied to the electronic device, and includes the following steps: acquiring, through an auxiliary unit, a target image request to determine or generate an image processing instruction (S10); processing, according to the image processing instruction, image data through an image processing unit to obtain a target image (S20). In the method, the target image request is acquired through the auxiliary unit, the image processing instruction is determined or generated according to the target image request; the image data is processed according to the image processing instruction and through the image processing unit, to obtain the target image. A new computational photography system architecture including the auxiliary unit and the image processing unit is provided, capable of achieving computational photography image processing.