Xuejuan Hu

In this study, we introduced a novel dual‐laser multi‐color imaging system. Integrated with a multi‐channel filter wheel, this system compared three spectral decontamination algorithms (nonnegative matrix factorization [NMF], RCAN, and PICASSO) showcasing its efficacy in achieving four‐color imaging with only two laser sources. Combined with a reliable image reconstruction algorithm, the spatial resolution of four channels super‐resolution four‐color images reached 130, 125, 133, and 132 nm, respectively. Lipid droplets, mitochondria, lysosomes, and nuclei from the mouse hepatocytes (AML12), human neuroblastoma cells (SH‐SY5Y), mouse hippocampal neuronal cells (HT‐22), and immortalized murine bone marrow‐derived macrophages were imaged. At the same time, the chromatin condensation, nuclear contraction, DNA fragmentation, apoptotic body formation, as well as the fusion of Mito and Lyso involved in mitochondrial autophagy were observed in HT‐22 and SH‐SY5Y cells suffering oxidative stress. Our multi‐color SIM imaging system establishes a powerful platform for dynamic organelle studies and other high‐resolution investigations in live cells.

Due to the intricate and diverse nature of textile defects, detecting them poses an exceptionally challenging task. In comparison with conventional defect detection methods, deep learning-based defect detection methods generally exhibit superior precision. However, utilizing deep learning for defect detection requires a substantial volume of training data, which can be particularly challenging to accumulate for textile flaws. To augment the fabric defect dataset and enhance fabric defect detection accuracy, we propose a fabric defect image generation method based on Pix2Pix generative adversarial network. This approach devises a novel dual-stage W-net generative adversarial network. By increasing the network depth, this model can effectively extract intricate textile image features, thereby enhancing its ability to expand information sharing capacity. The dual-stage W-net generative adversarial network allows generating desired defects on defect-free textile images. We conduct quality assessment of the generated fabric defect images resulting in peak signal-to-noise ratio and structural similarity values exceeding 30 and 0.930, respectively, and a learned perceptual image patch similarity value no greater than 0.085, demonstrating the effectiveness of fabric defect data augmentation. The effectiveness of dual-stage W-net generative adversarial network is established through multiple comparative experiments evaluating the generated images. By examining the detection performance before and after data augmentation, the results demonstrate that mean average precision improves by 6.13% and 14.57% on YOLO V5 and faster recurrent convolutional neural networks detection models, respectively.

Due to the periodicity of circuit boards, the registration algorithm based on keypoints is less robust in circuit board detection and is prone to misregistration problems. In this paper, the binary neighborhood coordinate descriptor (BNCD) is proposed and applied to circuit board image registration. The BNCD consists of three parts: neighborhood description, coordinate description, and brightness description. The neighborhood description contains the grayscale information of the neighborhood, which is the main part of BNCD. The coordinate description introduces the actual position of the keypoints in the image, which solves the problem of inter-period matching of keypoints. The brightness description introduces the concept of bright and dark points, which improves the distinguishability of BNCD and reduces the calculation amount of matching. Experimental results show that in circuit board image registration, the matching precision rate and recall rate of BNCD is better than that of classic algorithms such as scale-invariant feature transform (SIFT) and speeded up robust features (SURF), and the calculation of descriptors takes less time.

Visualizing the dynamics of fine structures in cells requires noninvasive, long-duration imaging of the intracellular environment at high spatiotemporal resolution and low background noise. With modularized hardware and polarized laser modulation, we developed a nanoscale super-resolution structured illumination microscope (SR-SIM) imaging technique. Combined with a reliable image reconstruction algorithm and timing synchronization of all devices, the super-resolution (SR) images in hepatocytes reached 134.95 nm spatial resolution and 50 fps temporal resolution. This imaging system was able to maintain the optimal operation state over hundreds of time points due to less exposure and low phototoxicity. In hepatocytes, interactions between mitochondria and lipid droplets (LDs) underpin many crucial physiological processes, ranging from cellular metabolism to signaling. In this study, we pioneered the use of the SIM system for imaging the interaction characteristics between mitochondria and LDs. More than 200 hepatocytes were counted and recorded effectively. We found that LDs in an unstable state were divided under mitochondrial contact and fused without it. Among 200 LDs, more than 69% were surrounded by mitochondria that tended to wrap LDs. The SR-SIM imaging technique was demonstrated to break the limitations of conventional imaging methods in spatial-temporal resolution and imaging duration in the field of the dynamic study between mitochondria and LDs.

In this paper, a structured illumination microscopy (SIM) image reconstruction algorithm combined with notch function (N-SIM) is proposed. This method suppresses the defocus signal in the imaging process by processing the low-frequency signal of the image. The existing super-resolution image reconstruction algorithm produces streak artifacts caused by defocus signal. The experimental results show that the algorithm proposed in our study can well suppress the streak artifacts caused by defocused signals during the imaging process without losing the effective information of the image. The image reconstruction algorithm is used to analyze the mouse hepatocytes, and the image processing tool developed by MATLAB is applied to identify, detect and count the reconstructed images of mitochondria and lipid droplets, respectively. It is found that the mitochondrial activity in oxidative stress induced growth inhibitor 1 (OSGIN1) overexpressed mouse hepatocytes is higher than that in normal cells, and the interaction with lipid droplets is more obvious. This paper provides a reliable subcellular observation platform, which is very meaningful for biomedical work.

Light-emitting diode (LED) chips have disordered arrangement and defects with characteristics of low contrast, for which traditional segmentation methods cannot classify surface defects effectively. In this paper, a chip segmentation method based on position pre-estimation and a modified Normalized Correlation Coefficient (NCC) matching algorithm, as well as feature enhancement methods are proposed. The position pre-estimation method is used to avoid the interference introduced by the disordered chip arrangement and the large missing area. By modifying the NCC algorithm, matching speed is improved by eight times compared to traditional NCC while matching result is not affected by brightness change. Furthermore, feature enhancement schemes with higher speed and accuracy were designed to identify low-contrast defects. The experimental results showed that the average accuracy reached 99.54%, improved by 0.66% compared to the state-of-the-art method while the inspection missing rate was 0.03%. In addition, the detection time of a single chip was approximately 1.098 ms, which meets the requirements of online detection, and the smallest defect that could be detected was 2 µm. In summary, the methods proposed in this study meet the requirements of industrial online detection regardless of accuracy, efficiency, or extensibility.