Handwritten Document Search Using Stroke and Character Code Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for retrieving handwritten documents are inefficient, as they do not effectively utilize the unique features of handwritten characters or figures, especially when searched by users other than the creator, leading to difficulties in accurately finding specific documents among many stored.
Innovation Solution
An information processing apparatus with a storage processor and search module that stores handwritten document data as time-series information, allowing for both handwriting searches based on stroke features and character searches using character codes, enabling precise retrieval of documents by utilizing unique user-specific features and character recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If handwritten documents are stored with unique user features, then search accuracy for the creator is improved, but searchability for other users deteriorates
Solution Approach 1:
The patent segments the feature extraction process into two distinct parts: user-specific features (for accurate retrieval by creators) and universal features (for cross-user searchability). The feature extraction unit separately identifies and stores these feature types, allowing the search system to appropriately weight each based on the query context, thereby resolving the contradiction between creator-specific accuracy and general searchability
Solution Approach 2:
The patent changes the parameter of feature representation by extracting multiple types of features with different characteristics. User-specific parameters (stroke order, shape variations) are combined with universal parameters (character recognition, semantic meaning), allowing the system to adapt the search strategy based on whether the query is from a creator or another user, thus balancing both requirements
2Ease of operation
If only character recognition is used for search, then ease of operation is improved, but measurement precision deteriorates
Solution Approach 1:
The patent merges multiple search approaches into a unified system: character recognition-based search (for ease of operation and universal understanding) is combined with stroke feature-based search (for high precision and user-specific accuracy). The search unit integrates both methods and can weigh them appropriately based on the query type, achieving both operational ease and search precision simultaneously
3Measurement precision
If stroke features are used for search, then search precision is improved, but device complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-extracting and pre-storing both user-specific features and universal features during the document creation phase. The feature extraction unit processes stroke data and identifies characteristic features in advance, organizing them in the storage unit for efficient retrieval. This preliminary processing reduces the complexity of real-time search operations while maintaining high precision
Data Source
AI summary
According to one embodiment, an information processing apparatus includes a storage processor and a search module. The storage processor stores document data and character codes, the document data including stroke data corresponding to strokes input by a handwriting operation and the character codes corresponding to the stroke data. The search module performs at least one of a handwriting search according to strokes of a first search key and a character search according to a character code of a second search key, stroke data corresponding to the strokes of the first search key retrieved from the document data in the handwriting search and stroke data corresponding to the character code of the second search key retrieved from the character codes in the character search.