Time-Lapse Quantitative Analysis of Drying Patterns and Machine Learning for Classifying Abnormalities in Sessile Blood Droplets

Anusuya Pal,Miho Yanagisawa,Amalesh Gope

DOI: https://doi.org/10.1101/2024.05.15.24307398

2024-05-17

Abstract:When a colloidal droplet dries on a substrate, a unique pattern results from multi-facet phenomena such as Marangoni convection, capillary flow, mass transport, mechanical stress, colloid-colloid, and colloid-substrate interactions. Even under uniform conditions (surface wettability, humidity, and temperature), slight differences in the initial colloidal composition alter the drying pattern. This paper shows how the evolving patterns during drying in the sessile droplets depend on the initial composition and are crucial for assessing any abnormalities in the blood. To do so, texture statistics are derived from time-lapse images acquired during drying, and different traditional machine learning are applied. In addition, a neural network analysis is performed on both images and their texture statistics. As the drying phenomena are correlated with the varying composition, these methods exhibit excellent performance in distinguishing blood abnormalities with an F1 score of over 97%. This indicates that analysis of time-lapse images during drying and their texture statistics, rather than conventional analysis using images at the final dry state, are crucial for classification. Our results highlight the potential of droplet drying as a low-volume, accurate, and simple screening tool for detecting the type and stage of any disease in bio-fluid samples, such as blood, urine, and saliva.

What problem does this paper attempt to address?

### Problems the paper attempts to solve

This paper aims to identify and classify abnormalities in blood by analyzing the patterns formed during the drying process of blood droplets. Specifically, the paper focuses on the following aspects:

1. **Pattern evolution during the drying process**:
- Research how the patterns formed during the drying process of blood droplets change with different initial compositions. These patterns are affected by multiple factors, such as Marangoni convection, capillary flow, mass transfer, mechanical stress, colloid - colloid interactions, and colloid - substrate interactions.

2. **Time - series image analysis**:
- Use time - series images to obtain texture features at different stages during the drying process. These features include first - order statistics (such as mean, standard deviation, skewness, kurtosis) and gray - level co - occurrence matrices (such as angular second - moment, contrast, correlation, inverse difference moment, entropy).

3. **Application of machine learning and neural networks**:
- Apply traditional machine - learning algorithms (such as decision trees, random forests, support vector machines, K - nearest neighbors, naive Bayes) and neural networks (such as artificial neural networks, convolutional neural networks) to classify time - series images and their texture features in order to distinguish between normal blood and blood containing abnormal components.

4. **Development of diagnostic tools**:
- Through the above methods, the paper shows how to use time - series images and texture features during the drying process to achieve high - precision classification of blood abnormalities. The research results indicate that this method has potential application value in detecting disease types and stages in biological fluid samples (such as blood, urine, saliva).

### Main contributions

- **Innovative analysis methods**: Combining time - series images, quantitative image - texture analysis, and data - driven methods improves the accuracy of classification.
- **Low - volume, high - precision screening tools**: By analyzing the drying patterns of a small amount of liquid, it provides a simple, fast, and accurate disease - screening method.
- **Wide applicability**: This method is not only applicable to blood but can also be extended to the analysis of other biological fluid samples and has broad application prospects.

### Conclusion

This paper successfully identifies and classifies abnormalities in blood through detailed time - series image analysis and machine - learning techniques. The research results show that time - series images and texture features during the drying process have important application value for the early diagnosis of diseases.