Search String Construction for Item Use Image Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users, especially non-reading individuals, face difficulties in creating effective search strings for image retrieval as they often lack the necessary words or phrases to describe item usage, leading to unsatisfactory results, such as images of product packaging instead of usage illustrations.
Innovation Solution
A computer-implemented method constructs, evaluates, and improves search strings by forming tuples of item classes, actions, and actors, using ontology siblings and similarity scores to generate alternative search strings that effectively indicate item use, providing more meaningful image retrieval results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users manually create search strings to retrieve images, then search results can be precise, but users without reading ability or vocabulary knowledge cannot effectively create search strings
Solution Approach 1:
The system automatically generates search strings and retrieves images without requiring user input or vocabulary knowledge. The processing device autonomously creates tuples from item descriptions, constructs appropriate search strings, and executes image retrieval operations, making the service accessible to users regardless of reading ability or search query formulation skills
Solution Approach 2:
The system introduces an intermediary processing layer that translates item descriptions into search strings and retrieves images. This intermediary layer handles the complex tasks of tuple construction, search string generation, and image retrieval automatically, shielding users from these technical complexities while delivering precise results
2Reliability
If reverse image searching is used to find similar images, then duplicate content and source images can be identified, but images indicating common item use cannot be effectively retrieved
Solution Approach 1:
The system segments the image retrieval task into distinct phases: first retrieving images based on visual similarity using reverse image search, then filtering and evaluating results to identify images showing common item use. This segmentation allows the system to leverage reverse image search for initial matching while adding specialized filtering for usage context
Solution Approach 2:
The system performs preliminary reverse image search to retrieve candidate images, then applies additional evaluation criteria to filter for images indicating common use. This preliminary retrieval followed by targeted filtering ensures both the comprehensiveness of visual matching and the specificity of usage-related results
3Productivity
If search strings focus on item names, then packaging images can be retrieved, but images illustrating actual item usage cannot be obtained
Solution Approach 1:
The system transitions from one-dimensional search (item name only) to multi-dimensional search by incorporating usage context, common verbs, and action-oriented terms into search strings. This dimensional expansion enables retrieval of images that capture both item identification and usage scenario, addressing the loss of usage context information
Solution Approach 2:
The system changes the parameters of search string construction by incorporating not just item names but also common verbs associated with item usage, action descriptors, and contextual terms. These parameter changes in search string composition enable the retrieval system to distinguish between packaging images and actual usage illustrations
Data Source
AI summary
Examples of techniques for constructing, evaluating, and improving a search string for retrieving images are disclosed. In one example implementation according to aspects of the present disclosure, a computer-implemented method includes constructing, by a processing device, a search string based at least in part on a tuple including an item class, an action, and an actor. The method further includes retrieving, by the processing device, a plurality of images based at least in part on the search string for an item. The method further includes evaluating, by the processing device, the retrieved plurality of images based on a similarity to determine whether the search string is effective at indicating a common item use. The method further includes, based at least in part on determining that the search string is ineffective at indicating the item use, generating, by the processing device, an alternative search string.