Phase unwrapping is a critical step to obtaining a continuous phase distribution in optical phase measurements and coherent imaging techniques. Traditional phase-unwrapping methods are generally low performance due to significant noise or undersampling. This paper proposes a deep convolutional neural network (DCNN) with a weighted jump-edge attention mechanism, namely, VDE-Net, to realize effective and robust phase unwrapping. Experimental results revealed that the weighted jump-edge attention mechanism, which is first proposed and simple to calculate, is useful for phase unwrapping. The proposed algorithm outperformed other networks or common attention mechanisms. In addition, an unseen wrapped phase image of a living red blood cell (RBC) was successfully unwrapped by the trained VDE-Net, thereby demonstrating its strong generalization capability.

Balanced dispersion between reference and sample arms is critical in frequency-domain optical coherence tomography (FD-OCT) to perform imaging with the optimal axial resolution, and the spectroscopic analysis of each voxel in FD-OCT can provide the metric of the spectrogram. Here we revisited dispersion mismatch in the spectrogram view using the spectroscopic analysis of voxels in FD-OCT and uncovered that the dispersion mismatch disturbs the A-scan’s spectrogram and reshapes the depth-resolved spectra in the spectrogram. Based on this spectroscopic effect of dispersion mismatch on A-scan’s spectrogram, we proposed a numerical method to detect dispersion mismatch and perform dispersion compensation for FD-OCT. The proposed method can visually and quantitatively detect and compensate for dispersion mismatch in FD-OCT, with visualization, high sensitivity, and independence from sample structures. Experimental results of tape and mouse eye suggest that this technique can be an effective method for the detection and compensation of dispersion mismatch in FD-OCT.

Automatic identification of fungi in microscopic fecal images provides important information for evaluating digestive diseases. To date, disease diagnosis is primarily performed by manual techniques. However, the accuracy of this approach depends on the operator’s expertise and subjective factors. The proposed system automatically identifies fungi in microscopic fecal images that contain other cells and impurities under complex environments. We segment images twice to obtain the correct area of interest, and select ten features, including the circle number, concavity point, and other basic features, to filter fungi. An artificial neural network (ANN) system is used to identify the fungi. The first stage (ANN-1) processes features from five images in differing focal lengths; the second stage (ANN-2) identifies the fungi using the ANN-1 output values. Images in differing focal lengths can be used to improve the identification result. The system output accurately detects the image, whether or not it has fungi. If the image does have fungi, the system output counts the number of different fungi types.

Fecal samples can easily be collected and are representative of a person’s current health state; therefore, the demand for routine fecal examination has increased sharply. However, manual operation may pollute the samples, and low efficiency limits the general examination speed; therefore, automatic analysis is needed. Nevertheless, recognition exhaustion time and accuracy remain major challenges in automatic testing. Here, we introduce a fast and efficient cell-detection algorithm based on the Faster-R-CNN technique: the Resnet-152 convolutional neural network architecture. Additionally, a region proposal network and a network combined with principal component analysis are proposed for cell location and recognition in microscopic images. Our algorithm achieved a mean average precision of 84% and a 723 ms detection time per sample for 40,560 fecal images. Thus, this approach may provide a solid theoretical basis for real-time detection in routine clinical examinations while accelerating the process to satisfy increasing demand.

The microcharacter recognition on the distributed feedback (DFB) laser chip is critically essential but a challenging task for the quality control in the incoming chip inspection and optical device manufacturing lines. Hence, this article proposes, designs, and implements a micro-optical character recognition system based on the microscopic imaging system and deep convolutional neural network to facilitate the microcharacter recognition of DFB chips. The proposed micron optical character recognition (MOCR) contains four components, namely, a microscopic imaging system-based visual-based measurement (VBM), character detection, orientation correction, and character recognition. In particular, data augmentation methods considering limited datasets and some chip character features are studied so as to accommodate complex industrial environments and tiny DFB chips. Experiments from the DFB image dataset of a real optical device manufacturing line showed that it reached the accuracy of 98.8%, which took a high speed of over 12.8 fps coping with DFB chip images, and saved 72% of the labor in the inspection position compared to traditional methods. To the best of our knowledge, there are no existing reports focused on MOCR on DFB chips. So, the first realization and successful application of MOCR played a key role in the intelligent manufacturing and quality traceability of DFB chips and devices in the field of optical communications, which is also of great significance for further accelerating the development of the industrial Internet.

The analysis of fecal-type components for clinical diagnosis is important. The main examination involves the counting of red blood cells (RBCs), white blood cells (WBCs), and molds under the microscopic. With the development of machine vision, some vision-based detection schemes have been proposed. However, these methods have a single target for detection, with low detection efficiency and low accuracy. We proposed an algorithm to identify the visible image of fecal composition based on intelligent deep learning. The algorithm mainly includes region proposal and candidate recognition. In the process of segmentation, we proposed a morphology extraction algorithm in a complex background. As for the candidate recognition, we proposed a new convolutional neural network (CNN) architecture based on Inception-v3 and principal component analysis (PCA). This method achieves high-average Precision of 90.7%, which is better than the other mainstream CNN models. Finally, the images within the rectangle marks were obtained. The total time for detection of an image was roughly 1200 ms. The algorithm proposed in the present paper can be integrated into an automatic fecal detection system.