Off-axis Digital Holograms Research Articles

Object image reconstruction: method for reconstructing images from digital off-axis holograms using a generative adversarial network

Object image reconstruction: method for reconstructing images from digital off-axis holograms using a generative adversarial network

Fast and accurate phase processing in off-axis digital holography combining adaptive spatial filtering and an embedded GPU platform

Abstract Parallel-phase processing enables rapid phase extraction from off-axis digital holograms. To achieve fast and accurate results, the phase reconstruction processes were parallelized using improved filter algorithms and optimized programming strategies. First, an adaptive filtering method based on the Chan-Vese (CV) model which better suits parallelism was designed to extract the +1 term spectrum. We selected suitable computer unified device architecture (CUDA) libraries according to the characteristics of the key phase reconstruction steps. Acceleration technologies, such as virtual memory and shared memory, were used to improve the computational efficiency. Furthermore, we combined an improved 4f optical imaging system with an embedded graphic processing unit (GPU) platform to design a low-cost phase reconstruction system for off-axis digital holography. To verify the feasibility of our method, the reconstructed quality of the CV filtering method was estimated, and the run times of phase retrieval on the central processing unit (CPU) and embedded GPU were compared for off-axis holograms with different pixel sizes. Additionally, the dynamic fluctuation phase maps of water droplet evaporation were retrieved to demonstrate the real-time capability of the method.

Accurate and fast autofocusing in off-axis digital holography based on step reduction search and particle swarm optimization

The accurate reconstruction distance is one of the important parameters in digital holography. Typical autofocusing methods have been proposed for digital holography to determine the focusing plane. In these methods, a series of holograms at fixed-step intervals always needed to be reconstructed and then locate the focal plane by using an evaluation function, a procedure that is very time-consuming. In this paper, an accurate and fast autofocusing method for off-axis digital holograms is proposed. The method first employs the step reduction search algorithm and the integral amplitude modulus (AMP) evaluation function with low computational effort to iteratively reduce the searching range of the focal plane. Then the particle swarm optimization (PSO) algorithm and a difference-in-amplitude (DIF) evaluation function with high sensitivity are used to accurately locate the focal plane. Numerical and experimental results show that the proposed method significantly improves the focusing plane detection accuracy and computational efficiency compared with the typical autofocusing methods.

Reconstructing images of objects: method for reconstructing images from digital off-axis holograms based on a generative adversarial neural network

The reconstruction of object images that are located in 3D scene cross-sections using digital holography is described. The potential of generative adversarial networks for reconstructing cross-sections of 3D scenes composed of multiple layers of off-axis objects from holograms is investigated. Such scenes consist of a series of sections with objects that are not aligned with the camera’s axis. Digital holograms were used to reconstruct images of cross-sectional views of 3D scenes. It has been shown that the use of neural networks increases the speed and reconstruction quality, and reduces the image noise. A method for reconstructing images of objects using digital off-axis holograms and a generative adversarial neural network is proposed. The proposed method was tested on both numerically simulated and experimentally captured digital holograms. It was able to successfully reconstruct up to 8 cross-sections of a 3D scene from a single hologram. It was obtained that an average structural similarity index measure was equal to at least 0.73. Based on optically registered holograms, the method allowed us to reconstruct object image cross-sections of a 3D scene with a structural similarity index measure over cross-sections of a 3D scene of equal to 0.83. Therefore, the proposed technique provides the possibility for high-quality object image reconstruction and could be utilized in the analysis of micro- and macroobjects, including medical and biological applications, metrology, characterization of materials, surfaces, and volume media.

A method for suppressing spectrum aliasing of off-axis digital holograms based on Kronecker interpolation and backpropagation

The research proposes a new method to solve the spectrum aliasing and improve the resolution of off-axis digital holograms based on Kronecker interpolation and backpropagation algorithms. The method suppresses the direct current (DC) image and increase the low-frequency information with Kronecker interpolation of the hologram. The size of the Fresnel reconstructed image can be controlled with the reconstruction distance, and this feature is used to select the appropriate reconstruction distance for the Fresnel amplitude reconstruction of interpolated holograms to separate the real and imaginary images. To obtain the complete first-order spectral information for the object, the amplitude of the desired image is obtained with spatial domain filtering and then inverted. Finally, the amplitude and phase of the object are reconstructed according to the angular spectrum reconstruction algorithm. The results show that the method can solve the spectrum aliasing problem of the holograms and thus improve the resolution of off-axis holograms. In addition, the background noise of the reconstructed phase is reduced because the influence of the zero-order spectrum can be completely avoided.

Carrier-frequency estimation for digital holograms of phase objects.

Accurate estimation of carrier fringe frequency is essential for the demodulation of off-axis digital holograms. The fringe frequency is often associated with the amplitude peak of the cross-term in the two-dimensional Fourier transform of a digital hologram. We point out that this definition of carrier frequency is not valid in general for holograms associated with phase objects. We examine the carrier-envelope representation for digital holograms from the viewpoint of Mandel's criterion [J. Opt. Soc. Am.57, 613 (1967)10.1364/JOSA.57.000613]. An appropriate definition of carrier frequency is observed to be the centroid of the power spectrum associated with the cross term. This definition is shown to apply uniformly to holograms associated with phase objects, is robust to noise, and leads to the smoothest (or least fluctuating) envelope representation for the demodulated object wave. The proposed definition is illustrated with simulated as well as experimentally recorded off-axis holograms.

Single-frame reconstruction by fractional Fourier-transform domain filtering in off-axis digital holographic microscopy

We propose a single-frame zero-order-eliminated reconstruction method by fractional Fourier transform filtering for an off-axis digital hologram. The filtering in the fractional Fourier transform domain of the hologram can effectively improve the reconstruction resolution, but it is required to remove its zero-order term. With the zero-order-term elimination of the Laplacian hologram, the higher reconstruction resolution of a single-frame hologram is achieved by zero-padding the hologram and choosing the optimal option of the fractional-order number. The results demonstrate that the resolutions of reconstructed amplitude and phase images are obviously improved. It will have a promising application in real-time imaging for biological cells and moving objects.

Simultaneous dual-wavelength digital holographic microscopy as a tool for the analysis of keratoacanthoma skin samples

A keratoacanthoma (KA) skin tumor is usually caused by sun exposure and may be an alert sign prior to the development of a more aggressive tumor or skin cancer. Studying the shape of the KA cells and their 3D rendering visualization are important parameters to prevent its evolution to higher stages of tumor cells or skin cancer. KA cells shape can be obtained through digital holographic microscopy; for that purpose, a setup with two illumination wavelengths (532 and 638 nm) is implemented to render a synthetic wavelength of 3.2 μm that avoids wrapping the optical phase of the processed holograms and increases measurement range. To recover the optical phase, two off-axis digital holograms are simultaneously recorded at each wavelength. From the processed hologram height variations, the shape and length of KA cells, as well as the stratum corneum epidermal layer, are obtained as phase images. The results achieved aid to discriminate healthy from malignant cells.

Quantitative phase imaging of living red blood cells combining digital holographic microscopy and deep learning.

Digital holographic microscopy as a non-contacting, non-invasive, and highly accurate measurement technology, is becoming a valuable method for quantitatively investigating cells and tissues. Reconstruction of phases from a digital hologram is a key step in quantitative phase imaging for biological and biomedical research. This study proposes a two-stage deep convolutional neural network named VY-Net, to realize the effective and robust phase reconstruction of living red blood cells. The VY-Net can obtain the phase information of an object directly from a single-shot off-axis digital hologram. We also propose two new indices to evaluate the reconstructed phases. In experiments, the mean of the structural similarity index of reconstructed phases can reach 0.9309, and the mean of the accuracy of reconstructions of reconstructed phases is as high as 91.54%. An unseen phase map of a living human white blood cell is successfully reconstructed by the trained VY-Net, demonstrating its strong generality.

3D reconstruction of unstained weakly scattering cells from a single defocused hologram.

We investigate the problem of 3D complex field reconstruction corresponding to unstained red blood cells (RBCs) with a single defocused off-axis digital hologram. The main challenge in this problem is the localization of cells to the correct axial range. While investigating the volume recovery problem for a continuous phase object like the RBC, we observe an interesting feature of the backpropagated field that it does not show a clear focusing effect. Therefore, sparsity enforcement within the iterative optimization framework using a single hologram data frame cannot effectively restrict the reconstruction to the true object volume. For phase objects, it is known that the amplitude contrast of the backpropagated object field at the focus plane is minimum. We use this information available in the recovered object field in the hologram plane to device depth-dependent weights that are proportional to the inverse of amplitude contrast. This weight function is employed in the iterative steps of the optimization algorithm to assist the object volume localization. The overall reconstruction process is performed using the mean gradient descent (MGD) framework. Experimental illustrations of 3D volume reconstruction of the healthy as well as malaria-infected RBCs are presented. A test sample of polystyrene microsphere bead is also used to validate the axial localization capability of the proposed iterative technique. The proposed methodology is simple to implement experimentally and provides an approximate tomographic solution, which is axially restricted and consistent with the object field data.

Single capture bright field and off-axis digital holographic microscopy.

We report on a single capture approach for simultaneous incoherent bright field (BF) and laser-based quantitative phase imaging (QPI). Common-path digital holographic microscopy (DHM) is implemented in parallel with BF imaging within the optical path of a commercial optical microscope to achieve spatially multiplexed recording of white light images and digital off-axis holograms, which are subsequently numerically demultiplexed. The performance of the proposed multimodal concept is firstly determined by investigations on microspheres. Then, the application for label-free dual-mode QPI and BF imaging of living pancreatic tumor cells is demonstrated.

Zero-order term suppression in off-axis holography based on deep learning method

Zero-order image is the main source of noise in the hologram acquisition process. In this paper, a preprocessing method based on deep learning are proposed. Generative adversarial network is applied to suppress zero-order image of off-axis digital holograms. To simplify data preparation process, we propose a digital hologram corresponding database generation method. The simulated data is used to replace the experimental data in the training network. After testing, the experimental data can be processed into a hologram without zero-order image successfully. And the different types of images are reconstructed with high quality. This method shows high efficiency and excellent robustness.

Two-wavelength digital holography through fog

Interferometric detection enables the acquisition of the amplitude and phase of the optical field. By making use of the synthetic wavelength as a computational construct arising from digital processing of two off-axis digital holograms, it is possible to identify the shape of an object obscured by fog and further increase the imaging range due to the increased sensitivity in coherent detection. Experiments have been conducted inside a 27 m long fog tube filled with ultrasonically generated fog. We show the improved capabilities of synthetic phase imaging through fog and compare this technique with conventional active laser illumination imaging.

An interference iterative reconstruction algorithm based on a single off-axis digital hologram

Compared with in-line digital holography, off-axis digital holography can be used to image more kinds of objects without the limitation of weak phase and sparsity, but the resolution is lower. In this paper, we put forward a high-resolution iterative reconstruction algorithm for off-axis holography which requires only the recording of a single off-axis hologram. Interference and inverse interference processes are used to extract the object wave from the off-axis hologram. A frequency domain filter with increasing radius is used to improve the competitive advantage of low frequency information in the early iterations, while retaining the high frequency information. The constraints of total variation and positive absorption are implemented at the object plane to enhance the robustness of the algorithm. Numerical simulations and experiments show that the resolution of our method is higher than traditional off-axis digital holography. In addition, it is suitable for phase image when the object has a strong phase fluctuation, as well as for imaging dynamical objects.

Common-path off-axis single-pixel holographic imaging.

Common-path off-axis single-pixel holographic imaging (COSHI) is proposed to obtain complex amplitude information using an in-line interferometer and a single-pixel (point-like) detector. COSHI is more robust to disturbances such as vibration than the conventional single-pixel digital holography technique because of its common-path configuration. In addition, the number of measurements can be reduced due to COSHI's reconstruction process based on the Fourier fringe analysis. In COSHI, an off-axis digital hologram can be obtained using the structured patterns composed of Hadamard basis patterns and stationary tilted phase distribution. Interestingly, COSHI's space bandwidth is larger than of the conventional off-axis digital holography because COSHI does not reconstruct the self-correlation term of an object. The proposed method is theoretically confirmed and numerical and experimental results show its feasibility.

Complex object wave retrieval using single-pixel compressive imaging of an off-axis Fresnel hologram

A complex wavefield imaging using single-pixel compressive sensing of an off-axis digital Fresnel hologram is proposed. A digital micromirror device (DMD) sampling prior to the single-pixel detection does the sparse representation of the hologram. The digital hologram is computationally recovered using a compressive sensing (CS) algorithm, and then the complex object wavefield is reconstructed using a complex wave retrieval method. The impact of the use of two measurement operators namely random and Walsh–Hadamard measurement matrices on the reconstruction quality of the hologram and complex object wavefield is compared. The cases of CS algorithm using different sparsification criteria i.e. Haar wavelet, Fresnelet sparsification and without sparsification of the digital hologram are analysed. The computer simulation results of the proposed CS method reveals that the Walsh–Hadamard sampling with Fresnelet sparsification achieves good quality reconstructions of both intensity and phase of the object wavefield.

Adaptive reconstruction imaging based on K-means clustering in off-axis digital holography

An adaptive filtering method for off-axis digital holographic reconstruction is presented. The spatial spectrum distribution of an off-axis digital hologram is clustered based on the K-means clustering algorithm of unsupervised machine learning . The spatial-spectrum filtering can be carried out by the quantitative comparison of different filtering interception windows. The first-order spectrum including the object information can be automatically located and intercepted after clustering the spatial spectrum distribution of the hologram. The experiment results demonstrate that the reconstruction images with better resolution and imaging quality can be achieved by the adaptive filtering algorithm. This adaptive filtering reconstruction based on K-means clustering provides a new way of automatic reconstruction imaging for off-axis digital holography.

Optimal Fresnelet sparsification for compressive complex wave retrieval from an off-axis digital Fresnel hologram

Complex wave retrieval (CWR) from an off-axis digital Fresnel hologram proposed by Liebling et al. is a single-exposure scheme in which a non-linear holographic process is modeled using simultaneous linear equations. This linear model fits well into the compressive sensing (CS) framework. We propose a method in which the CS is applied to a CWR method with Fresnelet sparsification. This approach compensates for the noise in the retrieved Fresnel field that is aroused due to the linear modeling approximations involved in the reconstruction process. As a result, a superior intensity and phase reconstruction were obtained by this method. Four methods including the direct CWR (conventional) method, compressive CWR method without sparsification, compressive CWR method with Haar wavelet, and compressive CWR method were compared with Fresnelet sparsification. The simulation and the optical experimental results converge to the same conclusion: the Fresnelet sparsified case is the best choice for quantitative phase imaging using compressive off-axis digital Fresnel holography.

Comparative analysis of off-axis digital hologram binarization by error diffusion

A digital micromirror device (DMD) provides 2D- and 3D-scene reconstruction by displaying diffractive and holographic optical elements. The highest frame rates (tens of thousands of Hz) can be achieved if the displayed optical elements are binarized. Except for DMD applications, hologram binarization is useful for display creation, image encryption, information compression and storage, fast 3D printing, etc. Error diffusion is one of the most qualitative implementations of hologram binarization. In this paper, three group's weighting matrices of error diffusion are analyzed: 16 standard, eight dot and six diagonal matrices. Thus, 30 error diffusion methods were used for optically recorded off-axis digital hologram binarization. The quality of the image reconstruction from the binarized holograms was compared. Direct applications of error diffusion with large weighting matrices and dot diffusion provide the highest reconstruction quality. Seven metrics were used as the error diffusion threshold. Twelve bypass directions of error diffusion were analyzed. In addition, the joint application of Otsu threshold and complex bypass directions allows us to improve the quality of hologram binarization by 15%.

Single-frame reconstruction for improvement of off-axis digital holographic imaging based on image interpolation.

We present a method of single-frame reconstruction in off-axis digital holography with Kronecker-product interpolation to maximize spatial-frequency extraction. The Kronecker-product interpolation operated on a one-frame off-axis digital hologram can generate several aliasing spectra in its Fourier domain, where the zero-order aliasing spectra in the extrapolation region can be suppressed effectively. After adding all aliasing spectrum regions and substituting each aliasing spectrum in place with their sum, spectrum interception with a larger scope for filtering can be made in one of the aliasing spectrum regions. The experimental results demonstrate that the method is valid to improve the resolution in off-axis digital holography.