Active Learning Image Search via Relevance Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users of mobile electronic devices face inefficiencies in searching and sorting through large collections of digital images, as existing methods require laborious labeling and sorting processes to identify relevant images, which is inconvenient and unproductive.
Innovation Solution
A method using active learning on mobile devices to automatically label and group images by learning a query concept based on content and context information, utilizing relevance feedback to sample and refine images, thereby minimizing manual effort and enhancing search efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If keyword based search engine is utilized to identify relevant images, then search capability is improved, but manual labeling effort is still required
Solution Approach 1:
The system performs self-labeling by automatically analyzing image content through computer vision algorithms, extracting features, and generating labels without human intervention. The image processing unit autonomously categorizes images based on detected objects, scenes, and attributes, eliminating the need for manual labeling while maintaining search functionality.
Solution Approach 2:
The patent replaces the mechanical manual labeling process with automated computer vision and machine learning systems. Image recognition algorithms, feature extraction techniques, and automated tagging mechanisms substitute human operators, transforming the labeling task from a manual operation to an automated computational process.
2Ease of operation
If images are sorted by labeling or placing in folders, then organization is improved, but convenience deteriorates due to laborious process
Solution Approach 1:
The image management system automatically organizes images by analyzing content, detecting objects and scenes, and assigning appropriate categories and folders without user intervention. The system self-manages the organization task by generating metadata, creating hierarchical structures, and maintaining sorted arrangements based on learned patterns from image analysis.
Solution Approach 2:
The system performs preliminary organization by pre-processing images upon ingestion, automatically extracting features, generating labels, and placing images in appropriate folders before the user needs to access them. This advance organization eliminates the need for subsequent manual sorting operations.
3Measurement precision
If manual labeling of each image is performed, then search accuracy is improved, but time consumption increases significantly
Solution Approach 1:
The patent substitutes manual labeling with automated computer vision systems that analyze image content, detect objects and attributes, and generate accurate labels through machine learning models. The system replaces human visual inspection and decision-making with automated algorithms that process images rapidly while maintaining high accuracy through feature extraction and pattern recognition.
Solution Approach 2:
The system changes the parameters of the labeling process by transitioning from manual human operations to automated computational operations. This involves changing the processing speed parameter, accuracy parameter, and resource consumption parameter by employing sophisticated image analysis algorithms, deep learning models, and parallel processing techniques to achieve both speed and precision.
Data Source
AI summary
The present disclosure provides a method, an electronic device, and a user interface for searching a plurality of relevant via active searching. At first, a query image having a first subject and a second subject may be obtained from stored images. Next, a query concept based on content information of the query image may be learned by sampling first sample images from the stored images according to content information of the query image for relevance feedback. Then, the query concept is refined based on context information of the first selected images among the first sample images. Based on the refined query concept, relevant images among the stored images may be searched according to the query concept and grouped into a collection album.