Dynamic Footwear Size Recommendation Using 3D Profiling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing solutions for recommending footwear size, particularly for children, are inaccurate, cumbersome, and fail to dynamically account for foot growth and variability in shoe sizes across different manufacturers, leading to health issues and high return rates in shoe sales.
Innovation Solution
A computer-implemented method using 3D foot profiling and machine learning to predict foot growth patterns, recommend suitable footwear sizes, and alert users when a size change is needed, incorporating user profiles and shoe manufacturer data for personalized and dynamic recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple pictures are captured using front facing camera to measure footwear size, then footwear size can be measured, but the process becomes very cumbersome for the end user
Solution Approach 1:
The patent uses a single reference image containing pre-defined measurement markers and scaling information. Instead of capturing multiple pictures of the foot, the system copies the foot onto this standardized reference template, allowing automated measurement extraction without requiring multiple captures or complex image processing of different angles.
Solution Approach 2:
The patent introduces a reference object with known dimensions and measurement markers as an intermediary between the user's foot and the measurement system. This reference object serves as a mediator that provides scaling information and measurement grid, enabling accurate measurements from a single image without requiring multiple pictures or complex calibration procedures.
2Reliability
If pre-calculated growth curves based on statistics analysis are used to predict foot growth, then growth prediction can be provided, but the mechanism is based on static data and does not dynamically update
Solution Approach 1:
The patent transitions from static pre-calculated growth curves to a dynamic machine learning model that continuously learns and updates foot growth patterns. The system adapts to individual user characteristics and updates predictions in real-time based on new measurement data, making the growth prediction mechanism dynamic rather than static.
Solution Approach 2:
The patent implements a feedback loop where new foot measurement data from users is continuously fed back into the machine learning model. This feedback mechanism allows the system to learn from actual growth patterns and improve prediction accuracy over time, rather than relying on fixed historical statistics that cannot adapt to new information.
3Adaptability or versatility
If existing solutions recommend footwear size based on previous purchases, then some personalization is provided, but they do not provide accurate personalized recommendations and fail to focus on children's footwear
Solution Approach 1:
The patent performs preliminary foot scanning and measurement to capture accurate 3D foot geometry data before making any footwear recommendations. By obtaining precise baseline measurements of foot length, width, volume, and shape characteristics in advance, the system can provide accurate personalized recommendations rather than relying on previous purchase history alone.
Solution Approach 2:
The patent analyzes multiple local foot parameters including foot length, width, volume, and shape characteristics rather than relying on a single overall metric. This localized analysis of different foot regions enables more precise footwear size and fit recommendations that account for the specific geometry and proportions of each user's foot.
4Measurement precision
If reference devices with marks are used to capture foot images, then foot length can be measured, but an operator with experience is required to position the feet, making it a manual process
Solution Approach 1:
The patent enables the system to automatically detect and measure foot parameters without requiring operator intervention for positioning. The machine learning model automatically identifies foot boundaries, orientation, and key measurement points from the captured image, making the system self-sufficient and eliminating the need for experienced operators to manually position feet against reference marks.
Solution Approach 2:
The patent replaces the mechanical reference device with physical marks and manual positioning requirements with an automated computer vision and machine learning system. Instead of using physical reference objects that require careful positioning, the system uses algorithms to automatically detect and measure foot geometry from images, substituting mechanical measurement methods with computational analysis.
Data Source
AI summary
A computer implemented method and system for dynamically recommending footwear size includes receiving a plurality of images of a foot of a user during a current scan of the user foot, analysing the plurality of images to generate a 3D foot profile of the user, determining a foot size of the user based on the 3D foot profile, predicting a foot growth pattern of the user based on the 3D foot profile, and a user profile, recommending one or more footwear size for the user based on the foot size, the predicted foot growth pattern, one or more footwear brands and models, and predicting a time of next scan of the foot based on the recommended footwear size and the predicted foot growth pattern.