Background: In the past few years, U-Net based U-shaped architecture and skip-connections have made incredible progress in the field of medical image segmentation. U2-Net achieves good performance in computer vision. However, in the medical image segmentation task, U2-Net with over nesting is easy to overfit. Purpose: A 2D network structure TransU2-Net combining transformer and a lighter weight U2-Net is proposed for automatic segmentation of brain tumor magnetic resonance image (MRI). Methods: The light-weight U2-Net architecture not only obtains multi-scale information but also reduces redundant feature extraction. Meanwhile, the transformer block embedded in the stacked convolutional layer obtains more global information; the transformer with skip-connection enhances spatial domain information representation. A new multi-scale feature map fusion strategy as a postprocessing method was proposed for better fusing high and low-dimensional spatial information. Results: Our proposed model TransU2-Net achieves better segmentation results, on the BraTS2021 dataset, our method achieves an average dice coefficient of 88.17%; Evaluation on the publicly available MSD dataset, we perform tumor evaluation, we achieve a dice coefficient of 74.69%; in addition to comparing the TransU2-Net results are compared with previously proposed 2D segmentation methods. Conclusions: We propose an automatic medical image segmentation method combining transformers and U2-Net, which has good performance and is of clinical importance. The experimental results show that the proposed method outperforms other 2D medical image segmentation methods.

Deep neural network approaches have made remarkable progress in many machine learning tasks. However, the latest research indicates that they are vulnerable to adversarial perturbations. An adversary can easily mislead the network models by adding well-designed perturbations to the input. The cause of the adversarial examples is unclear. Therefore, it is challenging to build a defense mechanism. In this paper, we propose an image-to-image translation model to defend against adversarial examples. The proposed model is based on a conditional generative adversarial network, which consists of a generator and a discriminator. The generator is used to eliminate adversarial perturbations in the input. The discriminator is used to distinguish generated data from original clean data to improve the training process. In other words, our approach can map the adversarial images to the clean images, which are then fed to the target deep learning model. The defense mechanism is independent of the target model, and the structure of the framework is universal. A series of experiments conducted on MNIST and CIFAR10 show that the proposed method can defend against multiple types of attacks while maintaining good performance.

This study proposes a novel lightweight grape detection method. First, the backbone network of our method is Uniformer, which captures long-range dependencies and further improves the feature extraction capability. Then, a Bi-directional Path Aggregation Network (BiPANet) is presented to fuse low-resolution feature maps with strong semantic information and high-resolution feature maps with detailed information. BiPANet is constructed by introducing a novel cross-layer feature enhancement strategy into the Path Aggregation Network, which fuses more feature information with a significant reduction in the number of parameters and computational complexity. To improve the localization accuracy of the optimal bounding boxes, a Reposition Non-Maximum Suppression (R-NMS) algorithm is further proposed in post-processing. The algorithm performs repositioning operations on the optimal bounding boxes by using the position information of the bounding boxes around the optimal bounding boxes. Experiments on the WGISD show that our method achieves 87.7% mAP, 88.6% precision, 78.3% recall, 83.1% F1 score, and 46 FPS. Compared with YOLOx, YOLOv4, YOLOv3, Faster R-CNN, SSD, and RetinaNet, the mAP of our method is increased by 0.8%, 1.7%, 3.5%, 21.4%, 2.5%, and 13.3%, respectively, and the FPS of our method is increased by 2, 8, 2, 26, 0, and 10, respectively. Similar conclusions can be obtained on another grape dataset. Encouraging experimental results show that our method can achieve better performance than other recognized detection methods in the grape detection tasks.