The data basis for blood cell analysis is formed by 3D holograms and the reconstructed phase images, which allow insights into the composition of the cell and its interior. The interpretation of these data requires the following steps:
List
Data Storage
When working with large amounts of data, it is essential to store them in a well-structured manner and to regulate access to them. In the CellFace project, several terabytes of DHM image data have already been produced and new measurements are constantly being added. These data must now be converted into a uniform format and stored in a database. These steps require the determination of the most suitable data storage solution for this sample and the necessary processing steps such as conversion and compression. Data management includes the selection of the file server for storing files, the database server, the database architecture as well as the backup strategy.
Segmentation
First, it must be determined which image areas contain relevant objects or structures. Subsequently, the relevant areas can be prepared for processing by sophisticated machine learning algorithms.
Labeling
As it is well known, modern learning algorithms require a large amount of training data. The task of building a suitable ground truth must be supported by the machine.
Feature Extraction
The main challenges here are the different object groups (leukocytes, erythrocytes, tissue cells and parasites), the strongly varying object size and the problem itself (e.g. malaria or leukemia detection). The aim is to find features that are universally and robustly applicable but do not lose their interpretability.
Classification
Depending on the question, e.g. the assignment to a blood cell subgroup, the detection of malaria pathogens or the determination of the viability of parasites, different statements have to be made. The decisions made in this step must be precise and comprehensible in order to use the software effectively.
CellPhaser Webservice
An intuitive machine learning toolbox will be build in the end of the project. The acceptance of a machine judgement is based on the transparency of the generating methodology and the linkability with existing forms of presentation.