HMind are intelligent decision making support systems based on the recent artificial intelligence technologies

Aibion Technologies develops intellegent systems for decision making support in pathology and oncology diagnostics

Our solutions are based on the state-of-the-art deep learning and computer vision algorithms

Our models are trained on the huge amount of whole slide images obtained from the medical laboratories

Our main goal is to provide pathologists with the intelligent tool which improves the accuracy and efficiency of the diagnostic process

We are developing our own solution - HMind. HMind is a platform which comprises separate intelligent modules for different tissue types and pathology areas. Neuroblastoma (HM001) is our first developed HMind module. It allows to extract quantitative and qualitative parameters from the digital slide copy. Using these features, pathologist can make a final diagnosis.

Neuroblastoma MODULE (HM001)

1. Sample extraction

Neuroblastoma MODULE (HM001)

2. Calculation of sample quantitative characteristics / extraction of information

Neuroblastoma MODULE (HM001)

3. Making a decision based on complete and comprehensive information about the sample

Neuroblastoma (HM001) module allows to define the histologic type of the tumor and to calculate quantitative parameters on the basis of the whole slide image.

The proportions of neuroblasts (differentiating and non-differentiated), ganglion cells (mature and immature), Schwann cells and neuropile are essential features for the tumor classification. Mitotic index is a tumor proliferation score which determines disease agressiveness and response to chemotherapy.

Aibion Technologies has developed AI solution which enables automatic detection and classification of seven cell types, neuropile and Schwann cels identification, automatic localization of mitotic or apoptotic cells. Currently the solution is being prepared for the clinical trials.


Better precision compared to the classical pathology diagnostics


The solution processes the entire whole slide image without skipping areas


Implementation of the module significantly reduces the whole slide analysis time