Multimode optical fiber plays an important role in endoscope miniaturization. With the development of deep learning and machine learning, neural networks can be used to identify and classify speckle patterns obtained at the fiber output. Based on the speckle pattern of a HERLEV dataset cell image transmitted by a multimode fiber, this paper studies the recognition accuracy of various types of speckle by a support vector machine, K-nearest neighbor, and convolutional neural network (Inception V3). Meanwhile, we propose an image classification optimization algorithm based on improved Inception V3. The experimental results show that the improved algorithm model is better than the traditional machine learning method; the accuracy rate is 97.92%, which effectively improves the performance of the pathological cell diagnosis deep learning model and lays a theoretical and practical foundation for further clinical application.

With the development of infrared technology, infrared dim and small target detection plays a vital role in precision guidance applications. To address the problems of insufficient dataset coverage and huge actual shooting costs in infrared dim and small target detection methods, this paper proposes a method for generating infrared dim and small target sequence datasets based on generative adversarial networks (GANs). Specifically, first, the improved deep convolutional generative adversarial network (DCGAN) model is used to generate clear images of the infrared sky background. Then, target–background sequence images are constructed using multi-scale feature extraction and improved conditional generative adversarial networks. This method fully considers the infrared characteristics of the target and the background, which can achieve effective expansion of the image data and provide a test set for the infrared small target detection and recognition algorithm. In addition, the classifier’s performance can be improved by expanding the training set, which enhances the accuracy and effect of infrared dim and small target detection based on deep learning. After experimental evaluation, the dataset generated by this method is similar to the real infrared dataset, and the model detection accuracy can be improved after training with the latest deep learning model.

Atmospheric turbulence is a common phenomenon in nature, in which the images obtained are often severely distorted, thus posing a significant challenge to the field of imaging. Computational ghost imaging (CGI), as an indirect imaging modality that exploits second-order correlation algorithms to reconstruct objects, exhibits a strong resistance to turbulence. However, constraints of sampling rate hamper its further application. To address this, data-driven deep learning methods have been proposed, demonstrating superior performance in image reconstruction in low-sampling conditions. While conventional data-driven deep learning approaches demonstrate strong task-specific performance, they are constrained by inherent limitations in generalizability and interpretability. In this paper, we propose a CGI method for atmospheric turbulence that integrates both model-driven and data-driven deep learning techniques. Unlike conventional deep learning methods, our approach combines these two strategies, leveraging the rich implicit features of data-driven methods alongside the generalization and interpretability advantages of model-driven approaches. Simulation and experimental results demonstrate that the proposed method is robust across various sampling ratios and turbulent conditions. Thus, our results provide an effective way for high-quality imaging in atmospheric turbulence.