A novel computer-aided detection method based on deep learning framework was proposed to detect small intestinal ulcer and erosion in wireless capsule endoscopy (WCE) images. To the best of our knowledge, this is the first time that deep learning framework has been exploited on automated ulcer and erosion detection in WCE images. Compared with the traditional detection method, deep learning framework can produce image features directly from the data and increase recognition accuracy as well as efficiency, especially for big data. The developed method included image cropping and image compression. The AlexNet convolutional neural network was trained to the database with tens of thousands of WCE images to differentiate lesion and normal tissue. The results of ulcer and erosion detection reached a high accuracy of 95.16% and 95.34%, sensitivity of 96.80% and 93.67%, and specificity of 94.79% and 95.98%, correspondingly. The area under the receiver operating characteristic curve was over 0.98 in both of the networks. The promising results indicate that the proposed method has the potential to work in tandem with doctors to efficiently detect intestinal ulcer and erosion.

Explainable artificial intelligence has been extensively studied recently. However, the research of interpretable methods in the agricultural field has not been systematically studied. We studied the interpretability of deep learning models in different agricultural classification tasks based on the fruit leaves dataset. The purpose is to explore whether the classification model is more inclined to extract the appearance characteristics of leaves or the texture characteristics of leaf lesions during the feature extraction process. The dataset was arranged into three experiments with different categories. In each experiment, the VGG, GoogLeNet, and ResNet models were used and the ResNet-attention model was applied with three interpretable methods. The results show that the ResNet model has the highest accuracy rate in the three experiments, which are 99.11%, 99.4%, and 99.89%, respectively. It is also found that the attention module could improve the feature extraction of the model, and clarify the focus of the model in different experiments when extracting features. These results will help agricultural practitioners better apply deep learning models to solve more practical problems.

As we know, most natural cells are transparent in intensity images. Thus, stains or fluorescent dyes have been used to help visualize them. But, for some cases, such as living cultured cells, it is different to stain them. Moreover, long-term observation on fluorescent sample will cause photobleaching and phototoxicity. Phase imaging can ease these problems. For example, with the differential interference contrast microscopy,¹ label-free living cells can be easily monitored. Nevertheless, this kind of method only records qualitative phase information. Digital holographic microscopy (DHM) can record quantitative phase information,^2,3 with high sensitivity, accuracy, and resolution, in a noninvasive way. But, in this method, the sample phase is often embedded within aberrations, highly limiting its applications. Fortunately, these aberrations can be quantified and then compensated with physical or numerical methods.^{4 – 8}

The multiphoton absorption (MPA) in MAPbBr ₃ and MAPbCl ₃ single-crystals (SCs) is experimentally investigated at the excitation wavelength from near-infrared to mid-infrared. A five-photon absorption (5PA) in MAPbBr ₃ is demonstrated at an excitation wavelength of 2400 nm for the first time, while that in MAPbCl ₃ demonstrated at 2100 nm via observing fluorescence as well as using Z-scan measurement. Meanwhile, a 5PA coefficient obtained is 8.4 × 10 ⁻¹³ cm ⁷ GW ⁻⁴ at 2400 nm for MAPbBr ₃ , while that of MAPbCl ₃ is 8.7 × 10 ⁻¹¹ cm ⁷ GW ⁻⁴ at 2100 nm. Our results reveal that mid-infrared is available wavelength for generating high-order MPA effect in MAPbBr ₃ and MAPbCl ₃ SCs.

A dual fan-shaped structure covered with Ag films was investigated for generating twin photonic hooks (t-PHs). The t-PH characteristics of this structure are studied using the Finite-Difference Time-Domain (FDTD) method. The results show that by designing appropriate fan-shaped opening angles and angles of Ag films coverage, the switching between t-PHs, S-shaped t-PHs, and W-shaped t-PHs can be achieved, along with controlling over the bending angles. The maximum first, second, and third bending angles for the obtained W-shaped t-PHs are 51.3°, 36.4°, and 41.8°, respectively, while the Ag films angle is 5°. The investigated tunable morphology t-PHs provide innovative applications in the fields of nanolithography and integrated optics.