A Dense Encoder–Decoder Network with Feedback Connections for Pan-Sharpening

To meet the need for multispectral images having high spatial resolution in practical applications, we propose a dense encoder–decoder network with feedback connections for pan-sharpening. Our network consists of four parts. The first part consists of two identical subnetworks, one each to extract features from PAN and MS images, respectively. The second part is an efficient feature-extraction block. We hope that the network can focus on features at different scales, so we propose innovative multiscale feature-extraction blocks that fully extract effective features from networks of various depths and widths by using three multiscale feature-extraction blocks and two long-jump connections. The third part is the feature fusion and recovery network. We are inspired by the work on U-Net network improvements to propose a brand new encoder network structure with dense connections that improves network performance through effective connections to encoders and decoders at different scales. The fourth part is a continuous feedback connection operation with overfeedback to refine shallow features, which enables the network to obtain better reconstruction capabilities earlier. To demonstrate the effectiveness of our method, we performed several experiments. Experiments on various satellite datasets show that the proposed method outperforms existing methods. Our results show significant improvements over those from other models in terms of the multiple-target index values used to measure the spectral quality and spatial details of the generated images.

MDCwFB: A Multilevel Dense Connection Network with Feedback Connections for Pansharpening

In most practical applications of remote sensing images, high-resolution multispectral images are needed. Pansharpening aims to generate high-resolution multispectral (MS) images from the input of high spatial resolution single-band panchromatic (PAN) images and low spatial resolution multispectral images. Inspired by the remarkable results of other researchers in pansharpening based on deep learning, we propose a multilevel dense connection network with a feedback connection. Our network consists of four parts. The first part consists of two identical subnetworks to extract features from PAN and MS images. The second part is a multilevel feature fusion and recovery network, which is used to fuse images in the feature domain and to encode and decode features at different levels so that the network can fully capture different levels of information. The third part is a continuous feedback operation, which refines shallow features by feedback. The fourth part is an image reconstruction network. High-quality images are recovered by making full use of multistage decoding features through dense connections. Experiments on different satellite datasets show that our proposed method is superior to existing methods, through subjective visual evaluation and objective evaluation indicators. Compared with the results of other models, our results achieve significant gains on the multiple objective index values used to measure the spectral quality and spatial details of the generated image, namely spectral angle mapper (SAM), relative global dimensional synthesis error (ERGAS), and structural similarity (SSIM).

Half-hour global solar radiation forecasting based on static and dynamic multivariate neural networks

Precise global solar radiation (GSR) measurements in a given location are very essential for designing and supervising solar energy systems. In the case of rarity or absence of these measurements, it is important to have a theoretical or empirical model to compute the GSR values. Therefore, the main goal of this work is to offer, to designers and engineers of solar energy systems, an appropriate and accurate way to predict the half-hour global solar radiation (HHGSR) time series from some available meteorological parameters (relative humidity, air temperature, wind speed, precipitation, and acquisition time vector in half-hour scale). For that purpose, two intelligent models are developed: the first one is a multivariate dynamic neural network with feedback connection, and the second is a multivariate static neural network. The database used to build these models was recorded in Agdal’s meteorological station in Marrakesh, Morocco, during the years of 2013 and 2014, and it was divided into two subsets. The first subset is used for training and validating the models, and the second subset is used for testing the efficiency and the robustness of the developed models. The obtained results, in terms of the statistical performance indicators, demonstrate the efficiency of the developed forecasting models to accurately predict the HHGSR parameter in the city of Marrakesh, Morocco.

Innate immunity mediator STING modulates nascent DNA metabolism at stalled forks in human cells

The cGAS/STING pathway, part of the innate immune response to foreign DNA, is known to be activated by cell's own DNA arising from the processing of the genome, including the excision of nascent DNA at arrested replication forks. We found STING activation to affect nascent DNA processing, suggesting a novel, unexpected feedback connection between the two events. Depletion of STING suppressed and re-expression of the protein in STING-deficient cells upregulated degradation of nascent DNA. Fork arrest was accompanied by the STING pathway activation, and a STING mutant that does not activate the pathway failed to upregulate nascent strand degradation. Consistent with this, cells expressing the STING mutant had a reduced level of RPA on parental and nascent DNA of arrested forks as well as a reduced CHK1 activation compared to the cells with wild type STING. Together our findings reveal a novel connection between replication stress and innate immunity.

Pseudo compact almost automorphy of neutral type Clifford-valued neural networks with mixed delays

<p style='text-indent:20px;'>We consider a class of neutral type Clifford-valued cellular neural networks with discrete delays and infinitely distributed delays. Unlike most previous studies on Clifford-valued neural networks, we assume that the self feedback connection weights of the networks are Clifford numbers rather than real numbers. In order to study the existence of <inline-formula><tex-math id="M1">\begin{document}$ (\mu, \nu) $\end{document}</tex-math></inline-formula>-pseudo compact almost automorphic solutions of the networks, we prove a composition theorem of <inline-formula><tex-math id="M2">\begin{document}$ (\mu, \nu) $\end{document}</tex-math></inline-formula>-pseudo compact almost automorphic functions with varying deviating arguments. Based on this composition theorem and the fixed point theorem, we establish the existence and the uniqueness of <inline-formula><tex-math id="M3">\begin{document}$ (\mu, \nu) $\end{document}</tex-math></inline-formula>-pseudo compact almost automorphic solutions of the networks. Then, we investigate the global exponential stability of the solution by employing differential inequality techniques. Finally, we give an example to illustrate our theoretical finding. Our results obtained in this paper are completely new, even when the considered networks are degenerated into real-valued, complex-valued or quaternion-valued networks.</p>

Initial Development of Mechanical Arm Support for Spinal Muscular Atrophy Patient

Spinal Muscular Atrophy or SMA disorder is generally thought to have affected as many as 1 in 40 in country like USA. SMA type 2 and 3 are commonly found in Indonesia. Children who suffer from SMA disease cannot move their hands with flexion - extension and abduction - adduction orientation maximally, because the muscles in the patient cannot support the movements of the hand. In this project, an initial joint effort was conducted by SGU and Bioteknik Design, to develop an active mechanical arm support with muscular feedback. Hence, the hand movements of the patient can be supported externally through detecting the muscle tension produced by moving the patient's hand. The tension is detected using Carbon Nanotube (CNT)-coated thread Mechanomyography (MMG) sensor. The sensor signal is then fed into Arduino microcontroller, to give appropriate control signal to the universal power window motor. As an early proof of concept, wood was used as the main structural material for the arm support. The project, however, did not go as expected due to the lack of torque from the motor and missed feedback connection from the sensor. A counter balance mechanism like spring may be attached for future improvement.

Effect of Top-Down Connections in Hierarchical Sparse Coding

Hierarchical sparse coding (HSC) is a powerful model to efficiently represent multidimensional, structured data such as images. The simplest solution to solve this computationally hard problem is to decompose it into independent layer-wise subproblems. However, neuroscientific evidence would suggest interconnecting these subproblems as in predictive coding (PC) theory, which adds top-down connections between consecutive layers. In this study, we introduce a new model, 2-layer sparse predictive coding (2L-SPC), to assess the impact of this interlayer feedback connection. In particular, the 2L-SPC is compared with a hierarchical Lasso (Hi-La) network made out of a sequence of independent Lasso layers. The 2L-SPC and a 2-layer Hi-La networks are trained on four different databases and with different sparsity parameters on each layer. First, we show that the overall prediction error generated by 2L-SPC is lower thanks to the feedback mechanism as it transfers prediction error between layers. Second, we demonstrate that the inference stage of the 2L-SPC is faster to converge and generates a refined representation in the second layer compared to the Hi-La model. Third, we show that the 2L-SPC top-down connection accelerates the learning process of the HSC problem. Finally, the analysis of the emerging dictionaries shows that the 2L-SPC features are more generic and present a larger spatial extension.

Predictive coding account of action perception: Evidence from effective connectivity in the Action Observation Network

Visual perception of actions is supported by a network of brain regions in the occipito-temporal, parietal, and premotor cortex in the primate brain, known as the Action Observation Network (AON). Although there is a growing body of research that characterizes the functional properties of each node of this network, the communication and direction of information flow between the nodes is unclear. According to the predictive coding account of action perception, this network is not a purely feedforward system but has feedback connections through which prediction error signals are communicated between the regions of the AON. In the present study, we investigated the effective connectivity of the AON in an experimental setting where the human subjects’ predictions about the observed agent were violated, using fMRI and Dynamical Causal Modeling (DCM). We specifically examined the influence of the lowest and highest nodes in the AON hierarchy, pSTS and ventral premotor cortex, respectively, on the middle node, inferior parietal cortex during prediction violation. Our DCM results suggest that the influence on the inferior parietal node is through a feedback connection from ventral premotor cortex during perception of actions that violate people’s predictions.