Pose Recognition Node Architecture for Interactive Graphic Effects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional pose estimation systems require extensive programming and are not flexible enough for modern graphics design, making it challenging for users to create interactive pose-based graphic effects that are reactive to human motion in images or videos.
Innovation Solution
A node architecture that uses a trained machine learning model to generate vector representations of human poses, allowing users to create interactive graphic effects without extensive coding, by recognizing poses and applying visual effects through a combination of recognizer nodes, compound nodes, data nodes, and render nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional pose estimation systems are used, then pose detection capability is achieved, but programming complexity and inflexibility increase
Solution Approach 1:
The patent introduces an intermediary system that translates between pose estimation output and graphics design input. This intermediary layer provides pre-built pose templates and effect templates that automatically map pose data to graphical effects, eliminating the need for users to write complex programming code while maintaining flexibility in graphics design.
Solution Approach 2:
The patent uses template copying where pre-defined pose templates and effect templates are created once and can be repeatedly applied to different videos. Users can copy existing templates and modify them without programming, enabling flexible graphics design while reducing complexity through reuse of proven configurations.
2Ease of operation
If extensive programming is required, then specific pose effects can be coded, but user accessibility and ease of use deteriorate
Solution Approach 1:
The patent implements dynamic template adjustment where users can visually modify pose templates and effect templates through drag-and-drop interfaces or parameter sliders without programming. The system dynamically updates the effect in real-time as users adjust parameters, making the system both easy to operate and highly customizable.
Solution Approach 2:
The patent provides pre-configured pose templates and effect templates that are ready to use immediately. Users can start with these preliminary configurations and make simple adjustments without needing to program from scratch, significantly improving ease of operation while maintaining full customization capability.
3Productivity
If traditional coding approaches are used, then pose-based effects can be implemented, but development time and complexity increase
Solution Approach 1:
The patent segments the effect creation process into independent, reusable templates: pose templates define pose detection parameters, effect templates define graphical output, and transition templates define animations between poses. This segmentation allows users to work with modular components rather than complex monolithic code, dramatically improving productivity while reducing perceived system complexity.
Solution Approach 2:
The patent creates universal templates that can be applied across different videos and use cases. A single pose template can recognize multiple poses, and a single effect template can work with different video inputs. This multi-functionality eliminates the need to create separate code for each effect, boosting productivity without increasing system complexity.
Data Source
AI summary
Embodiments are disclosed for interactive pose-based graphic effects. The method includes receiving an image including at least one object having a pose, the pose defined by an orientation and a position of the object within the image. A set of key joint data that represents the orientation and position of one or more points of interest associated with the object is generated. A vector representation of the set of key joint data is created for classifying one or more additional images that each include a candidate pose. One or more additional images are received. A match is detected between one of the candidate poses in the one or more additional images and the pose by comparing the vector representation of the set of key joint data to the candidate pose. A visual effect is generated based on the match.