Adaptive Cell Detection Using Peripheral State Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies face challenges in accurately determining whether a target cell is present in an image by refining candidates based on individual cell features and states, particularly for nucleated and anucleated red blood cells.
Innovation Solution
An image processing device that sets an object region, extracts cell regions, determines cell states, and sets feature quantity ranges for morphological features to identify target cells, using a learning unit to refine candidates based on positive and negative examples.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing uses fixed threshold conditions for target cell detection, then the processing is simple and fast, but the detection precision decreases when cell states vary (dense vs solitary)
Solution Approach 1:
The patent applies dynamics by making the detection threshold adaptive rather than fixed. The system dynamically adjusts the threshold based on the determined cell state (dense or solitary) in the peripheral region. When a solitary state is detected, a first threshold is applied; when a dense state is detected, a second threshold is applied. This dynamic adjustment resolves the contradiction by maintaining high detection precision across varying cell conditions while keeping the processing logic relatively simple through state-based threshold selection.
2Measurement precision
If the system processes only the object region for target cell detection, then the processing speed is fast, but the detection accuracy decreases due to lack of contextual information from surrounding cells
Solution Approach 1:
The patent applies the extraction principle by selectively processing only the necessary peripheral region around the object region, rather than processing the entire image. The system extracts cell state information from a limited peripheral area to inform threshold selection for the object region. This approach maintains fast processing by limiting the scope of additional analysis while still incorporating contextual information from surrounding cells to improve detection accuracy.
3Measurement precision
If the system uses state-dependent threshold selection based on peripheral cell analysis, then the detection precision is improved, but the processing complexity increases
Solution Approach 1:
The patent applies local quality by applying different processing strategies to different regions of the image. The object region receives threshold-based detection, while the peripheral region receives cell state analysis. Based on the peripheral region's cell state (dense or solitary), the system selectively applies appropriate thresholds to the object region. This localized approach improves detection precision by adapting to local conditions while managing overall complexity through region-specific processing.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image processing device 20 acquires captured images obtained by imaging a sample including a target cell, sets an object region for detecting the target cell in the captured images, extracts a plurality of cell regions each including a cell from a peripheral region including a region other than the object region, sets a range of a feature quantity of the target cell based on a feature quantity obtained from each of the plurality of cell regions, and determines whether or not the target cell is included in the object region when the feature quantity obtained from the object region is included in the set range of the feature quantity.