Learning Data Generation for Spinning Sphere Contour
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The accuracy of segmenting the contour of a sphere moving while spinning at high speed is compromised due to motion blur, making it difficult to determine where the sphere starts and ends.
Innovation Solution
A learning data generation device that includes a spinning rate estimation unit, a contour determination unit, and a learning data output unit, which processes a learning video image to estimate the spinning rate and determine the contour of the sphere with high accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a general segmentation tool is used to detect the contour of a spinning sphere, then the detection process is simple, but the segmentation accuracy is insufficient due to motion blur
Solution Approach 1:
The system performs preliminary actions by generating synthetic learning data with known ground truth contours and spinning rates before actual detection. This pre-training approach enables the recognition model to learn accurate contour estimation despite motion blur, resolving the contradiction between simple detection and high accuracy.
Solution Approach 2:
The system creates synthetic copies of spinning sphere videos with controlled parameters and known ground truth. These copied datasets are used to train the recognition model, enabling it to achieve high segmentation accuracy without requiring complex physical measurement setups.
2Measurement precision
If the spinning rate is not accurately estimated, then the contour determination is inaccurate, but obtaining the true spinning rate in advance requires complex external measurement equipment
Solution Approach 1:
The system performs self-service by estimating the spinning rate directly from the visual information in the video frames themselves, without requiring external measurement equipment. The recognition model learns to infer spinning rate from the patterns of motion blur and contour changes, enabling accurate measurement using only the camera system.
Solution Approach 2:
The system replaces mechanical spinning rate measurement equipment with a computational approach. Instead of using physical sensors or external devices to measure rotation, the system uses image processing and neural network-based estimation to derive spinning rate from visual data alone.
Data Source
AI summary
A learning data generation device for generating learning data for learning a recognizer capable of estimating a contour of a sphere making spinning motion, with high accuracy, the sphere being recorded in a single camera video image, is provided. The learning data generation device includes: a spinning rate estimation unit that receives an input of a learning video image in which motion of a spinning sphere is recorded and an initial value of a size of a contour of the recorded sphere in the video image, sets a plurality of set values of the size of the contour based on the initial value, and obtains an estimated value of a spinning rate of the sphere based on the learning video image, for each of the set values; a contour determination unit that receives an input of a true value of the spinning rate of the sphere, the true value being obtained in advance for the learning video image, and determines at least any of a plurality of the set values respectively corresponding to a plurality of the estimated values selected in order of closeness to the true value, as a determined value of the contour; and a learning data output unit that outputs the learning video image and the determined value as learning data.