Pose Estimation Smoothing for Jitter Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current pose estimation methods experience jitter and poor performance in following applications, leading to a floating feeling and unsatisfactory user experience.
Innovation Solution
A pose estimation method that acquires multiple frames of continuous reference images, processes rotation and translation pose components using specific smoothing algorithms, and generates a third pose estimation by combining smoothed rotation and translation components, thereby improving stability and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current pose estimation method is used, then pose estimation can be obtained, but jitter occurs and following performance is poor
Solution Approach 1:
The patent applies preliminary action by performing smoothing processing on rotation and translation pose components before generating the final pose estimation. Specifically, the method processes reference images and their pose components in advance to reduce jitter and improve stability, rather than simply processing the target image alone. This preliminary smoothing of temporal sequences resolves the contradiction by preparing stable pose data before it is needed for final output.
Solution Approach 2:
The patent implements feedback through the use of reference images that include both previous and next frames relative to the target image. The pose estimation system incorporates feedback from temporal sequences of images, using the previous frame's pose information and the next frame's pose information to smooth and refine the current frame's pose estimation. This feedback mechanism reduces jitter and improves following performance by continuously adjusting pose estimates based on temporal context.
2Productivity
If pose estimation is performed without smoothing, then processing is faster, but floating feeling occurs and user experience deteriorates
Solution Approach 1:
The patent applies local quality by treating rotation and translation pose components differently through separate smoothing processing. Instead of applying a uniform smoothing approach to all pose components, the method selectively processes rotation components and translation components with appropriate smoothing algorithms tailored to their specific characteristics. This localized approach maintains processing efficiency while significantly improving user experience by reducing floating feelings in the final pose estimation output.
Data Source
AI summary
A pose estimation method is provided. The method includes: according to timing sequence information, acquiring frames of continuous reference images before and after a target image on timing sequence; acquiring a first pose estimation of a target object in each frame of reference image, a second pose estimation of the target object in a target image; according to a rotation smoothing algorithm, processing a rotation pose component of at least one first pose estimation and a rotation pose component of the second pose estimation, to generate a target rotation pose component; according to a translation smoothing algorithm, processing a translation pose component of each first pose estimation and a translation pose component of the second pose estimation, to generate a target translation pose component; generating a third pose estimation of the target object in the target image according to the target rotation pose component and the target translation pose component.