Product Media File Clustering via Tile Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for clustering product media files in the ecommerce industry are manual and cumbersome, requiring significant time and resources, especially as the number of products increases, and often rely on manual file naming conventions that burden photo shoot teams and retailers.
Innovation Solution
A method using deep learning techniques to automatically cluster product media files by dividing them into tiles, computing feature vectors, and grouping similar tiles into patch clusters, which are then used to generate product groups without reliance on specific naming conventions, reducing the burden on photo shoot teams and marketers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual clustering methods are used to group product media files, then product groups can be generated, but the process becomes cumbersome and time-consuming as the number of products increases
Solution Approach 1:
The patent segments media files into multiple tiles and computes feature vectors for each tile independently. This segmentation allows parallel processing of numerous media files, dramatically reducing clustering time while maintaining accurate product grouping through comparative analysis of tile features across different media files
Solution Approach 2:
The patent replaces manual mechanical clustering processes with automated computer-based image processing and machine learning algorithms. The system automatically computes feature vectors, compares them across media files, and generates product groups without human intervention, eliminating the time-consuming manual sorting process while preserving grouping accuracy
2Ease of manufacture
If manual file naming conventions are required for product clustering, then product groups can be generated, but the burden on photo shoot teams and retailers increases
Solution Approach 1:
The patent implements self-service clustering where the system automatically identifies and groups media files belonging to the same product based on intrinsic visual features extracted from image tiles. This eliminates the need for external naming conventions or manual labeling, as the system autonomously performs clustering based on content analysis, reducing complexity for photo shoot teams and retailers
Solution Approach 2:
The patent extracts meaningful visual features directly from media file content through tile-based image processing and feature vector computation. By taking out and analyzing the essential visual characteristics of products, the system creates clusters based on actual product appearance rather than relying on external naming conventions, simplifying the overall process
Data Source
AI summary
A method for clustering product media files is provided. The method includes dividing each media file corresponding to one or more products into a plurality of tiles. The media file include one of an image or a video. Feature vectors are computed for each tile of each media file. One or more patch clusters are generated using the plurality of tiles. Each patch cluster includes tiles having feature vectors similar to each other. The feature vectors of each media file are compared with feature vectors of each patch cluster. Based on comparison, product groups are then generated. All media files having comparison output similar to each other are grouped into one product group. Each product group includes one or more media files for one product. Apparatus for substantially performing the method as described herein is also provided.