CELL VIABILITY ANALYSIS AND COUNTING FROM HOLOGRAMS BY USING DEEP LEARNING AND APPROPRIATE LENSLESS HOLOGRAPHIC MICROSCOPE

摘要

The invention is a holographic microscope (1) which detects the difference between the dead and live cells directly from the hologram images by training the deep learning based convolutional neural network and then makes predictions for viability analysis from the cell holograms obtained from the new samples (A) that were not used for training, and does not contain lens, mirror and similar optical elements, characterized in that, it comprises the following; a light source (10) which can be a laser or a light emitting diode (LED), an image sensor (30) which captures the images, a microfluidic chip (20) where the sample (A) located, a convolutional neural network which is formed in a server, is trained by predefining the hologram and/or phase images of dead and live cells, which are stained with Trypan blue or not and are stationary or flowing, and enable to make viability analysis to the samples (A).