Reference-Object Volume Prediction from Segmented Video Frames
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for calculating the volume of objects, such as video analysis, photographic measurement, and geometric modeling, face challenges in accuracy due to angle, lighting, texture, color fluctuations, and computational inefficiencies, especially when dealing with variable density foods.
Innovation Solution
A method using a computing device to predict the volume of a target object by obtaining a video, filtering frames based on blur detection, extracting object segmentation information, and calculating volume using multi-dimensional data, leveraging a reference object for accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If video analysis and photographic measurement are used to calculate object volume, then volume measurement can be performed, but measurement precision deteriorates due to angle and lighting requirements
Solution Approach 1:
The patent introduces a reference object as an intermediary element with known volume. By placing the reference object alongside the target object in the video, the system uses the reference object's known volume to calibrate and correct measurements of the target object, thereby improving measurement precision without requiring complex angle and lighting control
2Measurement precision
If geometric modeling is used to extract points from all objects in video, then volume prediction can be achieved, but computational time and processing amount increase excessively
Solution Approach 1:
The patent extracts only the necessary geometric information from the reference object and uses it to calculate the target object's volume through proportional relationships. Instead of processing all points from all objects in the video, the system extracts key dimensional data from the reference object and applies it to the target object, significantly reducing computational time while maintaining accuracy
Solution Approach 2:
The patent creates a simplified representation of the volume calculation by using the reference object as a template. The system copies the geometric modeling approach but applies it efficiently by using the reference object's known dimensions to scale and predict the target object's volume, avoiding the need for complex point-by-point analysis of all objects
3Adaptability or versatility
If geometric modeling is applied to all objects in video, then comprehensive volume analysis is possible, but device complexity increases
Solution Approach 1:
The patent segments the volume calculation task into two distinct parts: analyzing the reference object (which has known volume) and using its dimensional information to predict the target object's volume. This segmentation allows the system to maintain comprehensive volume analysis capability while reducing complexity by avoiding the need to process all objects in the video with full geometric modeling
Data Source
AI summary
Disclosed is a method for predicting a volume of an object, the method performed by one or more processors of a computing device according to an exemplary embodiment of the present disclosure.The method may includes: obtaining a video including an object and obtaining a plurality of frames included in the video; obtaining object segmentation information based on the plurality of frames; obtaining multi-dimensional data of the object based on the obtained object segmentation information; and calculating a volume of the object based on the multi-dimensional data of the object.