optical physics and artificial intelligence for diagnosis – OPAID

The OPAID project proposes the development of an original extended microscopy solution for reading blood smears and classifying circulating cells (white blood cells / WBC). The challenge is to overcome the limits of current solutions and avoid any return to manual reading under a microscope.
The proposed solution is based on a Fourier ptychography approach which produces an Optical Twin, a digital clone of the specimen that can be manipulated in a computational manner. The images produced are natively multimodal (intensity and phase) and have a resolution that exceeds the limits of conventional optics. The sharpness plane can be chosen a posteriori by calculation.
Based on initial preparatory work (TAMIS 2020-2023), the innovation concerns the equipment for data production and processing (UV source, etc.). The perspective aims for a major development for the detection and classification of peripheral blood WBCs on both coloured smears and uncoloured smears (label-free). The validation of the label-free will resolve the stability and survivorship problems of the reagents and will open up the possibility of reusing the slides for other explorations such as Raman spectrometry or immunocytochemistry.
Artificial Intelligence tools occupy a central place in the project and are applied both to the reconstruction of the images produced and to the various classification processes.
OPAID is supported by a well-established collaborative partnership between teams of biologists (APHP, Sorbonne University) specialists in haematological medical issues, the multi-disciplinary of Télécom SudParis combining mastery of Physical Optics and the exploitation of Artificial Intelligence, both for the reconstruction of computational imagery, the improvement of image quality and classification, and the company TRIBVN which has experience in both the development and deployment of imaging solutions diagnostic microscopy.