AI Video Calorie Estimation Using User Benchmarking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wearables and applications for calculating calories burned during activity performance are limited by non-reliance on individual variability, inaccurate baseline data, limited activity recognition, non-reliance on environmental factors, and lack of real-time feedback.
Innovation Solution
A method and system using Artificial Intelligence (AI) that processes real-time video streams and user profile attributes to determine activity parameters, select a target user for benchmarking, and calculate calories burned based on user efficiency levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If wearables use sensors and statistical information to estimate calories burnt, then the device complexity is reduced and battery life is conserved, but the measurement precision of calories burned is inaccurate
Solution Approach 1:
The patent introduces an intermediary AI-based calibration system that mediates between the simple sensor data collection and the need for precise calorie measurement. The system uses a calibration phase with known activity parameters to establish individual-specific models, which then enable accurate calorie calculations without requiring complex continuous monitoring hardware.
Solution Approach 2:
The patent creates a digital copy of the user's physiological response patterns through the calibration process. By capturing and storing the relationship between sensor data and actual calorie expenditure during calibration, the system generates a personalized model that replicates accurate calorie estimation without requiring complex real-time measurement hardware.
2Measurement precision
If wearables rely on agreed information and standard equations for calorie calculations, then the ease of operation is improved, but the measurement precision is reduced due to lack of individual variability consideration
Solution Approach 1:
The patent performs preliminary calibration action during an initial phase where the user completes activities with known parameters. This preliminary data collection and model building happens once during calibration, after which the system can automatically perform precise calorie measurements without requiring continuous user input or manual configuration during actual use.
3Loss of information
If wearables do not provide real-time feedback to conserve battery life, then the loss of energy is reduced, but the loss of information regarding activity performance is increased
Solution Approach 1:
The system performs all complex calculations and model updates during the preliminary calibration phase rather than continuously during use. This allows the system to provide rich real-time feedback during actual activities without the ongoing energy cost of complex processing, as the heavy computational work is completed beforehand during calibration.
4Measurement precision
If wearables use standard calorie equations without environmental factors, then the ease of manufacture is improved, but the measurement precision is reduced due to lack of environmental consideration
Solution Approach 1:
The system enables the user to provide environmental information voluntarily during the calibration process, and this information is automatically integrated into the personalized model without requiring continuous environmental monitoring or complex sensor arrays. The user's own observations and inputs during calibration serve to characterize their environmental context, which is then stored and applied to future measurements.
Data Source
AI summary
A method for calculating calories burned during activity performance using Artificial Intelligence (AI) is disclosed. The method includes receiving video stream of user performing activity and set of user profile attributes of the user. The method further includes creating, for each of the plurality of frames, multimedia vector corresponding to the associated frame that are further processed to determine set of activity parameters associated with user. The method includes selecting a target user from plurality of target users based on similarity between set of user profile attributes of the user and set of target profile attributes of the target user. The method further includes comparing each of the set of activity parameters with corresponding activity parameter from a set of target activity parameters associated with the target user. The method includes determining user efficiency level and count of calories burned by the user based on the result of comparing.