Figure 2
Source publication
Contexts in source publication
Context 1
... CWMF, ACWMF and so on. Their corresponding noise situation is also different. AMF is very adequate for salt and pepper noise, while MF is also apt for the noise above plus random noise. According to different failure types, different filter combinations are selected during pretreatment. This is the preprocessing part, the first step in the model (Fig. ...
Context 2
... above loss function is used to train the parameter Θ needed by the model. In this function, {(y i , x i )} í µí±í µí±=1 í µí±í µí± is N pairs of noisy and cleaning patches. Network Architecture: We set the depth of proposed model to 17, and different of layers are adopted in our model, as shown in Fig. 2. After pre-processing, we used one Conv plus ReLu layer to get a multi-channel result. That's because we use 64 convolution kernels, and they are all the same size of 3 × 3. So we can get 64 feature graphs in this layer. In order to make our results more robust, the activation function ReLU is used to make the results more nonlinear. ...
Similar publications
Het plicasyndroom is een klinische entiteit die wordt veroorzaakt door een ontstekingsproces
van een plica, een embryonaal residu, gewoonlijk in de knie, dat meestal
geen symptomen veroorzaakt. Het syndroom betreft vaker de mediale plica en slechts zelden
de suprapatellaire plica, nochtans de frequentste plica in de knie. Aanwezigheid van
een volle...
Citations
... The outcomes showed that the MaskRCNN method has revealed unsuccessful outcomes with the AP, PQ, SQ, and RQ of 30.44%, 50.71%, 61.6% and 65.11% correspondingly. After that, the Mask RCNN-COCO system has resulted in slightly enhanced AP, PQ, SQ and RQ values of 33.09%, 52.88%, 63% Finally, a detailed CDF examination of the PFPN-ADT model takes place on three datasets namely Cityscapes, ADE20k and Vistas under distinct Intersection Over Union (IOU) as shown in Tab. 4 and Fig. 10 [25][26][27]. The results depicted that the CDF values tend to increase with a rise in IOU. ...
... Artificial intelligence is widely used in various fields such as coverless information hiding techniques, machine vision and so on [35][36][37][38][39]. Medical images or signals can be analyzed with the assistance of AI to identify deviations from normal patterns that may indicate disease [40]. ...
... In order to achieve the best classification effect of XGBoost, we optimize parameters by crossvalidation. In the same experimental environment, we use the average absolute error MAE to measure the error between true label and predicted label [17][18][19]. The setting of parameters is shown in Tab ...
Spectrum sensing will be an essential component in developing cognitive radio networks, which will be an essential component of the subsequent generation of wireless communication systems. Over the course of several decades, a great deal of different strategies, including cyclo-stationary, energy detectors, and matching filters, have been put up as potential solutions. Obviously, each of these methods comes with a few of negatives that you have to take into consideration. When the Signal-to-Noise Ratio (SNR) changes, energy detectors work poorly; cyclo-stationary detectors are technically sophisticated; and employing matching filters needs experience with Primary User (PU) signals. Researchers have recently been devoting a great deal of attention to Machine Learning (ML) and Deep Learning (DL) algorithms as a result of the potential uses that these algorithms may have in the development of exceptionally accurate spectrum sensing models. The capacity to learn from data in a way that traditional learning algorithms are unable to has led to the rise in prominence of these types of algorithms. The Hybrid Model of Improved Long Short Term Memory with Improved Extreme Learning Models (HILSTM-IELM), to be more specific, is what is being suggested since it reduces the amount of energy that is used during data transmission as well as the range and the duty cycle. Because of this, the disadvantage in existing methodology, proposed technique reduced to a certain level in energy consumption. In the last step of this analysis, the performance of the HILSTM-IELM-based spectrum sensing is compared to that of a variety of different methods that are currently in use. According to the findings of recent studies, the spectrum sensing method that was created provides superior performance to that of technologies in terms of the accuracy, sensitivity, and specificity of data transmission systems.
This paper presents a multi-sequence satellite image removal algorithm to remove cloud cover in remote sensing images. The algorithm is based on dual residual network. Above all, the network uses multi-temporal cloud images as the input of the model. Then, it fuses the global and local feature information through the dual residual connection structure, which makes the overall structure of the generated restored image reasonable and the edge details clearer. Finally, it uses pixel up-sampling to enhance the utilization of spatial information and improve the restoration effect. Experiments on the European Sentinel-2 remote sensing satellite image dataset with different cloud removal algorithms demonstrate that our network achieves higher quality than the state-of-the-art methods. The proposed method was verified on the dataset, and the PSNR and SSIM indexes were 27.04 and 0.8479, both exceeding the original processing method STGAN of the dataset, and improving the effectiveness of remote sensing image cloud removal.
KeywordsImage cloud removalRemote sensing imageImage restorationDual Residuals NetworkMultitemporal images
In order to improve the performance of microphone array-based sound source localization (SSL), a robust SSL algorithm using convolutional neural network (CNN) is proposed in this paper. The Gammatone sub-band steered response power-phase transform (SRP-PHAT) spatial spectrum is adopted as the localization cue due to its feature correlation of consecutive sub-bands. Since CNN has the “weight sharing” characteristics and the advantage of processing tensor data, it is adopted to extract spatial location information from the localization cues. The Gammatone sub-band SRP-PHAT spatial spectrum are calculated through the microphone signals decomposed in frequency domain by Gammatone filters bank. The proposed algorithm takes a two-dimensional feature matrix which is assembled from Gammatone sub-band SRP-PHAT spatial spectrum within a frame as CNN input. Taking the advantage of powerful modeling capability of CNN, the two-dimensional feature matrices in diverse environments are used together to train the CNN model which reflects mapping regularity between the feature matrix and the azimuth of sound source. The estimated azimuth of the testing signal is predicted through the trained CNN model. Experimental results show the superiority of the proposed algorithm in SSL problem, it achieves significantly improved localization performance and capacity of robustness and generality in various acoustic environments.
