Search

Information extraction from synthetic aperture radar (SAR) images is heavily impaired by speckle noise, and hence, despeckling is a crucial preliminary step in scene analysis algorithms. The recent success of deep learning envisions a new generation of despeckling techniques that could outperform classical model-based methods. However, current deep learning approaches to despeckling require supervision for training, whereas clean SAR images are impossible to obtain. In the literature, this issue is tackled by resorting to either synthetically speckled optical images, which exhibit different properties with respect to true SAR images, or multitemporal SAR images, which are difficult to acquire or fuse accurately. In this article, inspired by recent works on blind-spot denoising networks, we propose a self-supervised Bayesian despeckling method. The proposed method is trained by employing only noisy SAR images and can, therefore, learn features of real SAR images rather than synthetic data. Experiments show that the performance of the proposed approach is very close to the supervised training approach on synthetic data and superior on real data in both quantitative and visual assessments. [ABSTRACT FROM AUTHOR]

Point clouds are an increasingly relevant geometric data type but they are often corrupted by noise and affected by the presence of outliers. We propose a deep learning method that can simultaneously denoise a point cloud and remove outliers in a single model. The core of the proposed method is a graph-convolutional neural network able to efficiently deal with the irregular domain and the permutation invariance problem typical of point clouds. The network is fully-convolutional and can build complex hierarchies of features by dynamically constructing neighborhood graphs from similarity among the high-dimensional feature representations of the points. The proposed approach outperforms state-of-the-art denoising methods showing robust performance in the challenging setup of high noise levels and in presence of structured noise. [ABSTRACT FROM AUTHOR]

Non-local self-similarity is well-known to be an effective prior for the image denoising problem. However, little work has been done to incorporate it in convolutional neural networks, which surpass non-local model-based methods despite only exploiting local information. In this paper, we propose a novel end-to-end trainable neural network architecture employing layers based on graph convolution operations, thereby creating neurons with non-local receptive fields. The graph convolution operation generalizes the classic convolution to arbitrary graphs. In this work, the graph is dynamically computed from similarities among the hidden features of the network, so that the powerful representation learning capabilities of the network are exploited to uncover self-similar patterns. We introduce a lightweight Edge-Conditioned Convolution which addresses vanishing gradient and over-parameterization issues of this particular graph convolution. Extensive experiments show state-of-the-art performance with improved qualitative and quantitative results on both synthetic Gaussian noise and real noise. [ABSTRACT FROM AUTHOR]

Recently, convolutional neural networks (CNNs) have been successfully applied to many remote sensing problems. However, deep learning techniques for multi-image super-resolution (SR) from multitemporal unregistered imagery have received little attention so far. This article proposes a novel CNN-based technique that exploits both spatial and temporal correlations to combine multiple images. This novel framework integrates the spatial registration task directly inside the CNN, and allows one to exploit the representation learning capabilities of the network to enhance registration accuracy. The entire SR process relies on a single CNN with three main stages: shared 2-D convolutions to extract high-dimensional features from the input images; a subnetwork proposing registration filters derived from the high-dimensional feature representations; 3-D convolutions for slow fusion of the features from multiple images. The whole network can be trained end-to-end to recover a single high-resolution image from multiple unregistered low-resolution images. The method presented in this article is the winner of the PROBA-V SR challenge issued by the European Space Agency (ESA). [ABSTRACT FROM AUTHOR]

Compressed sensing (CS) has recently emerged as an effective and efficient way to encrypt data. Under certain conditions, it has been shown to provide some secrecy notions. In theory, it could be considered to be a perfect match for constrained devices needing to acquire and protect the data with computationally cheap operations. However, the theoretical results on the secrecy of compressive cryptosystems only hold under the assumption of infinite precision representation. With this work, we aim to close this gap and lay the theoretical foundations to support this practical framework. We provide theoretical upper bounds on the distinguishability of the measurements acquired through finite precision sensing matrices and experimentally validate them. Our main result is that the secrecy of a CS cryptosystem can be exponentially increased with a linear increase in the representation precision. This result confirms that the CS can be an effective secrecy layer and provides tools to use it in practical settings. [ABSTRACT FROM AUTHOR]

Compression of hyperspectral images onboard of spacecrafts is a tradeoff between the limited computational resources and the ever-growing spatial and spectral resolution of the optical instruments. As such, it requires low-complexity algorithms with good rate-distortion performance and high throughput. In recent years, the Consultative Committee for Space Data Systems (CCSDS) has focused on lossless and near-lossless compression approaches based on predictive coding, resulting in the recently published CCSDS 123.0-B-2 recommended standard. While the in-loop reconstruction of quantized prediction residuals provides excellent rate-distortion performance for the near-lossless operating mode, it significantly constrains the achievable throughput due to data dependencies. In this paper, we study the performance of a faster method based on the prequantization of the image followed by a lossless predictive compressor. While this is well known to be suboptimal, one can exploit powerful signal models to reconstruct the image at the ground segment, recovering part of the suboptimality. In particular, we show that convolutional neural networks can be used for this task and that they can recover the whole SNR drop incurred at a bit rate of 2 bits per pixel. [ABSTRACT FROM AUTHOR]

This paper addresses the problem of training a deep neural network for satellite image segmentation so that it can be deployed over images whose statistics differ from those used for training. For example, in postdisaster damage assessment, the tight time constraints make it impractical to train a network from scratch for each image to be segmented. We propose a convolutional encoder–decoder network able to learn visual representations of increasing semantic level as its depth increases, allowing it to generalize over a wider range of satellite images. Then, we propose two additional methods to improve the network performance over each specific image to be segmented. First, we observe that updating the batch normalization layers’ statistics over the target image improves the network performance without human intervention. Second, we show that refining a trained network over a few samples of the image boosts the network performance with minimal human intervention. We evaluate our architecture over three data sets of satellite images, showing the state-of-the-art performance in binary segmentation of previously unseen images and competitive performance with respect to more complex techniques in a multiclass segmentation task. [ABSTRACT FROM AUTHOR]

Being able to reliably link a picture to the device that shot it is of paramount importance to give credit or assign responsibility to the author of the picture itself. However, this task needs to be performed at large scales due to the recent explosion in the number of photos taken and shared. Existing methods cannot satisfy those requirements. Methods based on the photo response nonuniformity (PRNU) of digital sensors are able to link a photo to the device that shot it and have already been used as proof in the Court of Law. Such methods are reliable but so far, they can be only used for small-scale forensic tasks involving few cameras and pictures. ToothPic, an acronym for "Who Took This Picture?," is a novel image retrieval engine that allows to find all the pictures in a large-scale database shot by a given query camera. [ABSTRACT FROM AUTHOR]