In the area of optical remote sensing image processing, object detection is among the utmost essential and difficult tasks. By virtue of the excellent feature representation abilities of deep convolutional neural networks (DCNNs), the performance of remote sensing object detection has recently increased dramatically. However, DCNN-based methods necessitate sufficient labeled training samples and tend to experience a considerable performance fall when training examples are insufficient. Many of the recently developed few-shot object detection (FSOD) methods attempt to solve the issue by using the idea of meta-learning, which intends to extract knowledge that can be generalized across various tasks. Despite its success, the negligence of the knowledge learned in the past, and inter-class correlations hinder the detection ability of novel classes. In this article, we propose a new meta-memory based FSOD approach named MM-RCNN. Specifically, MM-RCNN adopts a memory moduel to store each category’s knowledge learned in the training stage and the memory-based external attention (MEA) to aggregate all categories’ information simultaneously. Based on MEA, we design two feature enhancement modules for the region proposal network (RPN) and detection head to boost the performance. Experiments over two remote sensing benchmarks, i.e., DIOR and NWPU VHR10, verify the capability of our method (0.239 and 0.557 average mAP across all settings).

Vehicle detection in remote sensing images remains a challenge because most vehicles are small and cover only a relatively small area due to the low ground sample distance. Although image super-resolution can improve small object detection performance as a preprocessing step, methods for improving the quality of the entire image tend to focus on the majority backgrounds that are not important for detection and involve high computational cost. Inspired by the promising feature-level super-resolution method, in this letter, we propose a novel anchor-free vehicle detection network for small vehicle detection in remote sensing images. Specifically, a target-guided feature super-resolution network is proposed to enhance the features of the potential target. Besides, we propose a novel feature fusion module to improve the feature representation of shallow layers, which accounts for small object detection. Extensive experiments on three public remote sensing detection datasets [cars overhead with context (COWC), Vehicle Detection in Aerial Imagery (VEDAI), and UCAS-AOD] amply demonstrate that our method can achieve significant performance with a mean average precision of 0.933, 0.756, and 0.961, respectively.

Label noise may affect the generalization of classifiers, and the effective learning of main patterns from samples with noisy labels is an important challenge. Recent studies have shown that deep neural networks tend to prioritize the learning of simple patterns over the memorization of noise patterns. This suggests a possible method to search for the best generalization that learns the main pattern until the noise begins to be memorized. Traditional approaches often employ a clean validation set to find the best stop timing of learning, i.e., early stopping. However, the generalization performance of such methods relies on the quality of validation sets. Further, in practice, a clean validation set is sometimes difficult to obtain. To solve this problem, we propose a method that can estimate the optimal stopping timing without a clean validation set, called limited gradient descent. We modified the labels of a few samples in a noisy dataset to obtain false labels and to create a reverse pattern. By monitoring the learning progress of the noisy and reverse samples, we can determine the stop timing of learning. In this paper, we also theoretically provide some necessary conditions on learning with noisy labels. Experimental results on CIFAR-10 and CIFAR-100 datasets demonstrate that our approach has a comparable generalization performance to methods relying on a clean validation set. Thus, on the noisy Clothing-1M dataset, our approach surpasses methods that rely on a clean validation set.

Oriented object detection is a fundamental and challenging task in remote sensing image analysis that has recently drawn much attention. Currently, mainstream oriented object detectors are based on densely placed predefined anchors. However, the high number of anchors aggravates the positive and negative sample imbalance problem, which may lead to duplicate detections or missed detections. To address the problem, this paper proposes a novel anchor-free two-stage oriented object detector. We propose the Anchor-Free Oriented Region Proposal Network (AFO-RPN) to generate high-quality oriented proposals without enormous predefined anchors. To deal with rotation problems, we also propose a new representation of an oriented box based on a polar coordinate system. To solve the severe appearance ambiguity problems faced by anchor-free methods, we use a Criss-Cross Attention Feature Pyramid Network (CCA-FPN) to exploit the contextual information of each pixel and its neighbors in order to enhance the feature representation. Extensive experiments on three public remote sensing benchmarks—DOTA, DIOR-R, and HRSC2016—demonstrate that our method can achieve very promising detection performance, with a mean average precision (mAP) of 80.68%, 67.15%, and 90.45%, respectively, on the benchmarks.

Remote sensing images are widely used in many applications. However, due to being limited by the sensors, it is difficult to obtain high-resolution (HR) images from remote sensing images. In this paper, we propose a novel unsupervised cross-domain super-resolution method devoted to reconstructing a low-resolution (LR) remote sensing image guided by an unpaired HR visible natural image. Therefore, an unsupervised visible image-guided remote sensing image super-resolution network (UVRSR) is built. The network is divided into two learnable branches: a visible image-guided branch (VIG) and a remote sensing image-guided branch (RIG). As HR visible images can provide rich textures and sufficient high-frequency information, the purpose of VIG is to treat them as targets and make full use of their advantages in reconstruction. Specially, we first use a CycleGAN to drag the LR visible natural images to the remote sensing domain; then, we apply an SR network to upscale these simulated remote sensing domain LR images. However, the domain gap between SR remote sensing images and HR visible targets is massive. To enforce domain consistency, we propose a novel domain-ruled discriminator in the reconstruction. Furthermore, inspired by the zero-shot super-resolution network (ZSSR) to explore the internal information of remote sensing images, we add a remote sensing domain inner study to train the SR network in RIG. Sufficient experimental works show UVRSR can achieve superior results with state-of-the-art unpaired and remote sensing SR methods on several challenging remote sensing image datasets.

In this letter, we propose a novel solution to the problem of single image super-resolution at multiple scaling factors, including the extreme case of 8x, with a single network architecture. In applications where only a detail needs to be super-resolved, traditional solutions must choose to use as input either the low-resolution detail, thus losing the information about the context, or the whole low-resolution image and then crop the desired output detail, which is quite wasteful in terms of computations and storage. To address both of these issues we propose ZoomGAN, a model that takes as input the whole low-resolution image, which we call context, and a binary mask that specifies with a box which image detail in the low-resolution image to magnify. The output of ZoomGAN has the same size as the inputs so that the scaling factor is implicitly defined by the arbitrary size of the mask box. To encourage a realistic and high-quality output, we combine adversarial training with a perceptual loss. We use two discriminators: one promotes the similarity between the distributions of real and generated details and the other promotes the similarity between the distributions of real and generated (detail, context) pairs. We evaluate ZoomGAN with several experiments on several datasets and show that it achieves state of the art performance on zoomed in details in terms of the LPIPS and PI perceptual metrics, while being on par in terms of the PSNR distortion metric.