We propose an autofocusing algorithm to obtain, relatively accurately, the 3D position of each particle, particularly its axial location, and particle number of a dense transparent particle solution via its hologram. First, morphological analyses and constrained intensity are used on raw reconstructed images to obtain information on candidate focused particles. Second, axial resolution is used to obtain the real focused particles. Based on the mean intensity and equivalent diameter of each candidate focused particle, all focused particles are eventually secured. Our proposed method can rapidly provide relatively accurate ground-truth axial positions to solve the autofocusing problem that occurs with dense particles.

We propose a modified residual neural network (ResNet) to quickly and accurately detect the particle quantity from raw reconstructed images of a high-density particle solution in digital holography. The raw reconstructed image is fed into the modified ResNet to obtain the particle quantity. Then, the quantity and particle concentration in the captured volume are calculated. The metrological challenge is modeled as a regression problem in deep learning. The average relative error of the holograms in the test dataset is less than 10% even when predicting the particle quantities untrained by the model. Hence, an accurate particle quantity is obtained even when the raw reconstructed images are not denoised, thereby reducing the processing time.

We propose an autofocusing algorithm to obtain, relatively accurately, the 3D position of each particle, particularly its axial location, and particle number of a dense transparent particle solution via its hologram. First, morphological analyses and constrained intensity are used on raw reconstructed images to obtain information on candidate focused particles. Second, axial resolution is used to obtain the real focused particles. Based on the mean intensity and equivalent diameter of each candidate focused particle, all focused particles are eventually secured. Our proposed method can rapidly provide relatively accurate ground-truth axial positions to solve the autofocusing problem that occurs with dense particles.