Melanocytic lesions occur on the surface of the skin, in which the malignant type is melanoma with a high fatality rate, seriously endangering human health. The histopathological analysis is the gold standard for diagnosis of melanocytic lesions. In this study, a fully automated intelligent diagnosis method based on deep learning was proposed to classify the pathological whole slide images (WSI) of melanocytic lesions. Firstly, the color normalization based on CycleGAN neural network was performed on multi-center pathological WSI; Secondly, ResNet-152 neural network-based deep convolutional network prediction model was built using 745 WSI; Then, a decision fusion model was cascaded, which calculates the average prediction probability of each WSI; Finally, the diagnostic performance of the proposed method was verified by internal and external test sets containing 182 and 54 WSI, respectively. Experimental results showed that the overall diagnostic accuracy of the proposed method reached 94.12% in the internal test set and exceeded 90% in the external test set. Furthermore, the color normalization method adopted was superior to the traditional color statistics-based and staining separation-based methods in terms of structure preservation and artifact suppression. The results demonstrate that the proposed method can achieve high precision and strong robustness in pathological WSI classification of melanocytic lesions, which has the potential in promoting the clinical application of computer-aided pathological diagnosis.
Humans ; Deep Learning ; Melanoma/pathology* ; Diagnosis, Computer-Assisted ; Neural Networks, Computer ; Skin/pathology*

Recording the highly diverse and dynamic activities in large populations of neurons in behaving animals is crucial for a better understanding of how the brain works. To meet this challenge, extensive efforts have been devoted to developing functional fluorescent indicators and optical imaging techniques to optically monitor neural activity. Indeed, optical imaging potentially has extremely high throughput due to its non-invasive access to large brain regions and capability to sample neurons at high density, but the readout speed, such as the scanning speed in two-photon scanning microscopy, is often limited by various practical considerations. Among different imaging methods, light field microscopy features a highly parallelized 3D fluorescence imaging scheme and therefore promises a novel and faster strategy for functional imaging of neural activity. Here, we briefly review the working principles of various types of light field microscopes and their recent developments and applications in neuroscience studies. We also discuss strategies and considerations of optimizing light field microscopy for different experimental purposes, with illustrative examples in imaging zebrafish and mouse brains.
Animals ; Mice ; Microscopy/methods* ; Zebrafish ; Neurons/physiology* ; Brain/physiology* ; Neurosciences