Hierarchical Information Segmentation for Multi-Granularity Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information segmentation methods struggle to accurately determine demarcation points of varying granularities in information sequences, leading to inefficiencies in subsequent applications such as editing and recognition.
Innovation Solution
An information segmentation method that determines first and second demarcation point probabilities for a target information node, allowing for multiple segmentation modes with different granularities by using cascaded classifiers and advanced feature extraction networks to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing information segmentation methods are used, then the segmentation process is simple, but the accuracy of demarcation point determination is low
Solution Approach 1:
The patent applies segmentation by dividing the information sequence into multiple granularities through hierarchical segmentation. Different types of demarcation points (first type, second type, etc.) are identified at different hierarchical levels, allowing the system to segment information at varying granularities from fine to coarse levels, thereby improving measurement precision without overwhelming complexity
Solution Approach 2:
The patent changes parameters by computing multiple types of demarcation point probability values (first demarcation point probability, second demarcation point probability, etc.) for each information node. By varying the probability thresholds and types of demarcation points considered, the system achieves higher accuracy in determining segmentation points while maintaining a structured approach to complexity
2Adaptability or versatility
If multiple segmentation modes with different granularities are implemented, then the adaptability of the segmentation system is improved, but the computational complexity increases
Solution Approach 1:
The patent implements dynamics by providing multiple segmentation modes with different granularities that can be dynamically selected based on application needs. The system computes probability values for different types of demarcation points and allows flexible combination of these modes, enabling adaptability without requiring a completely separate system for each granularity level
Solution Approach 2:
The patent achieves universality by creating a unified segmentation framework that handles multiple types of demarcation points and multiple segmentation modes within a single system. The same computational infrastructure supports fine-grained and coarse-grained segmentation, reducing overall system complexity while maintaining high adaptability
3Manufacturing precision
If hierarchical segmentation from fine to coarse levels is performed, then the precision of information segmentation is improved, but the processing time increases
Solution Approach 1:
The patent applies preliminary action by computing probability values for different types of demarcation points in advance for all information nodes. These pre-computed probability values are then used during the actual segmentation process, allowing the system to make quick decisions about segmentation points without re-computing complex probabilities during segmentation, thus reducing processing time while maintaining high precision
Data Source
AI summary
Methods, apparatuses and electronic devices for information segmentation. The method includes, for a target information node in an information sequence and based on first node information of the target information node, determine a first demarcation point probability value, in which the first demarcation point probability value indicates a probability that the target information node is a first type of demarcation point; determine, based on the first demarcation point probability value and second node information of the target information node, a second demarcation point probability value, in which the second demarcation point probability value indicates a probability that the information node is a second type of demarcation point, and the first type of demarcation point is associated with the second type of demarcation point; and determines, based on the first and the second demarcation point probability value, at least two segmentation modes for the information sequence.