Holodetect Enables Rapid, AI-Enhanced, Fully Automated Cell Counting and Classification
Join us for our first appearance at AlgaEurope 2023 in Prague from 12-15 December. We will present our microscope designed to automate cell counting and classification using artificial intelligence. We look forward to connecting with potential partners in the field of algae research.
Cell counting by Holodetect requires minimal sample preparation.
Holodetect also provides accurate count estimates for cells in large conglomerates. Accuracy of the cell count has been validated against conventional counting methods.
Morphologic properties of different cell types and their distribution can also be measured and plotted.
Holodetect also combines Digital Holographic and Fluorescent Microscopy for higher accuracy and further measurement options.
By measuring fluorescence intensity on cell or sample level, the chlorophyll content of the cells and the sample can be estimated.
Based on their fluorescence levels, or using proper staining, living and dead cells can also be reliably distinguished by Holodetect.
Holodetect uses Deep Learning models for object focusing, segmentation and classification. These models can flexibly be retrained for classifying and counting various species.
Beside monocultures, cells in co-cultures can also be separately classified and counted.
Our few-shot learning module enables on-site retraining by the user in a matter of minutes
The measurement cuvettes used in Holodetect instruments - depending on the required magnification - are 200-10000 micron deep. Holodetect can take an image of the whole depth of the sample in a single snapshot using digital holographic microscopy with no need of refocusing. All objects in the sample can then be reconstructed by computer using the recorded hologram. This means minimal sample preparation, and enables using the microscope in flow-through mode and analyzing larger sample volumes.
The reconstructed images are analyzed using artificial intelligence. These AI models can be flexibly configured for various species and objects in the sample.
The holographic microscope can be combined with a fluorescent microscope. This can increase classification accuracy and enables various fluorescence related measurements on the sample.
Holodetect HR is a high resolution instrument that measures objects in the size range of 3 - 75 μm in flow-through mode with a measured volume of 3,6-18 μl/min and reports particle class, count, size and shape statistics.
This instrument performs cell classification and detection in mono- as well as multi species cultures, and provides reliable cell count estimates in conglomerates. Holodetect HR uses retrainable deep learning for object detection and classification. Report types can flexibly be configured.
Holodetect HR is shipped with or without built-in touch display.
Holodetect HiRes+F is a laboratory instrument for detecting and counting small particles or biological contents with fluorescence in fluids. In addition to Holodetect HiRes functions, Holodetect HiRes+F performs automatic cell counting based on the fluorescence of cells.
The fluorescence intensity is measured on cell or sample level. It can be used for approximating the chlorophyll content of the cells.
Using combined labeled and label-free fluorescence feature measurements HiRes+F can also reliably distinguish between living and dead cells.
Holodetect LVol is a large measured volume instrument that measures objects in the size range of 15 - 300 μm in flow-through mode with a measured volume of 10-50 ml/min and reports particle class, count, size and shape statistics. This instrument is well suited for larger cell size algae species (e.g. Spirulina). Holodetect LVol uses retrainable deep learning for object detection and classification. Report types can flexibly be configured.
Holodetect LVol can also be integrated into the control loop of the microalgae production process.
Holodetect instruments can be interfaced with existing industrial microalgae culturing solutions. We are ready to implement custom sampling, and measurement data relaying to the control loop in custom systems.
The AI models built in Holodetect can be flexibly trained and retrained to classify new species or objects in fluids. We would be happy to assist you in AI model design and training.
In order to successfully train deep learning models, a significant amount of annotated training data is required. We can help you to build those databases.
SZTAKI is a research institute with the fundamental task of performing basic and application-oriented research in an interdisciplinary setting in the fields of computer science, engineering, information technology, intelligent systems, process control, wide-area networking and multimedia.
Holodetect instruments are developed by the Computational Optical Sensing and Processing Laboratory of SZTAKI. One of its main research areas is the development of digital holographic and fluorescent microscopes with applications, among others, in water utilities, microalgae, and beer brewery industries.