Tensor completion aims at filling in the missing elements of an incomplete tensor based on its partial observations, which is a popular approach for image inpainting. Most existing methods for visual data recovery can...
详细信息
Tensor completion aims at filling in the missing elements of an incomplete tensor based on its partial observations, which is a popular approach for image inpainting. Most existing methods for visual data recovery can be categorized into traditional optimization-based and neural network-based methods. The former usually adopt a low-rank assumption to handle this ill-posed problem, enjoying good interpretability and generalization. However, as visual data are only approximately low rank, handcrafted low-rank priors may not capture the complex details properly, limiting the recovery performance. For neural network-based methods, despite their impressive performance in image inpainting, sufficient training data are required for parameter learning, and their generalization ability on the unseen data is a concern. In this paper, combining the advantages of these two distinct approaches, we propose a tensor Completion neural Network (CNet) for visual data completion. The CNet is comprised of two parts, namely, the encoder and decoder. The encoder is designed by exploiting the CANDECOMP/PARAFAC decomposition to produce a low-rank embedding of the target tensor, whose mechanism is interpretable. To compensate the drawback of the low-rank constraint, a decoder consisting of several convolutional layers is introduced to refine the low-rank embedding. The CNet only uses the observations of the incomplete tensor to recover its missing entries and thus is free from large training datasets. Extensive experiments in inpainting color images, grayscale video sequences, hyperspectral images, color video sequences, and light field images are conducted to showcase the superiority of CNet over state-of-the-art methods in terms of restoration performance.
In recent years, the accuracy of edge detection on several benchmarks has been significantly improved by deep learning based methods. However, the prediction of deep neural networks is usually blurry and needs further...
详细信息
In recent years, the accuracy of edge detection on several benchmarks has been significantly improved by deep learning based methods. However, the prediction of deep neural networks is usually blurry and needs further post-processing including non-maximum suppression and morphological thinning. In this paper, we demonstrate that the blurry effect arises from the binary cross-entropy loss, and crisp edges could be obtained directly from deep convolutional neural networks. We propose to learn edge maps as the representation of local contrast with a novel local contrast loss. The local contrast is optimized in a stochastic way to focus on specific edge directions. Experiments show that the edge detection network trained with local contrast loss achieves a high accuracy comparable to previous methods and dramatically improves the crispness. We also present several applications of the crisp edges, including image completion, image retrieval, sketch generation, and video stylization.
Hypercomplex signal and imageprocessing extends upon conventional methods by using hypercomplex numbers in a unified framework for algebra and geometry. The special issue is divided into two parts and is focused on c...
详细信息
Hypercomplex signal and imageprocessing extends upon conventional methods by using hypercomplex numbers in a unified framework for algebra and geometry. The special issue is divided into two parts and is focused on current advances and applications in computational signal and imageprocessing in the hypercomplex domain. The first part offered well-rounded coverage of the field, with seven articles that focused on overviews of current research, color imageprocessing, signal filtering, and machine learning.
The need for deep convolutional neural network is increasing for medical image classification because it provides good performance. This work elucidates the significance of convolutional neural network in making effec...
详细信息
The need for deep convolutional neural network is increasing for medical image classification because it provides good performance. This work elucidates the significance of convolutional neural network in making effective detection of clinical diseases by categorizing the clinical images in an organized manner. Clinical diseases are difficult to predict and interpret. To predict diseases from medical images, the stochastic Multinomial Logarithmic (SML) based image classification method is proposed. To effectively eliminate noise from images, edge-boosting locally adapted space-variant filters are first applied to the texture and medical MRI, and CT images. The SML approach is used to improve feature classification and disease prediction. Accuracy, Peak signal-to-Noise Ratio (PSNR), precision, recall and specificity performances of the proposed approach are compared with surviving methods. The proposed method produces enhanced performance compared to the existing ones with improved accuracies of 95.8% and 96.2 % respectively, for Brodatz texture and brain MRI, CT images.
Dynamic threshold neural P (DTNP) systems are neural-like computing models. This paper discusses how to use DTNP systems to propose a new change detection for SAR images. DTNP systems have two intrinsic and recognizab...
详细信息
Dynamic threshold neural P (DTNP) systems are neural-like computing models. This paper discusses how to use DTNP systems to propose a new change detection for SAR images. DTNP systems have two intrinsic and recognizable mechanisms: dynamic threshold and spiking mechanisms. The two machanisms are used to a new region growing algorithm. To obtain the optimal seed points of regional growth, particle swarm optimization (PSO) algorithmis used, and then DTNP systems are used to control the regional growth according to these seed points. Simulation experiments are carried out on four SAR image datasets, and the proposed method is evaluated on two metrics and compared with several state-of-the-art or baseline change detection methods. The comparison results show the effectiveness and advantage of the proposed method for change detection of SAR images.
image quality is significantly impacted by rain, posing challenges in fields like surveillance, autonomous driving, and outdoor robotics. The field of image deraining, particularly for single image, has attracted cons...
详细信息
image quality is significantly impacted by rain, posing challenges in fields like surveillance, autonomous driving, and outdoor robotics. The field of image deraining, particularly for single image, has attracted considerable attention to improve image clarity in inclement weather. To overcome the inherent complexity of the single-image rain removal task, we proposed a novel architecture of the attention mechanism and gated recurrent network (AMGR-Net) that combines spatial and channel attention mechanisms with gated recurrent units. AMGR-Net contains multiple modules, each of which uses convolution kernels and attention mechanisms to enhance feature extraction. AMGR-Net demonstrates superior performance over state-of-the-art methods in both synthetic and real-world image datasets, as evidenced by higher peak signal to noise ratio and structural similarity index measurement scores. The integration of spatial attention significantly enhances feature expression, enabling more effective rain streak removal and detail preservation. Furthermore, this method also shows promising results in the application of stripe noise removal from meteorological satellite cloud images. To overcome the inherent complexity of the single-image rain removal task, we proposed a novel architecture of the attention mechanism and gated recurrent network (AMGR-Net) that combines spatial and channel attention mechanisms with gated recurrent units. AMGR-Net contains multiple modules, each of which uses convolution kernels and attention mechanisms to enhance feature extraction. AMGR-Net demonstrates superior performance over state-of-the-art methods in both synthetic and real-world image datasets, as evidenced by higher peak signal to noise ratio and structural similarity index measurement scores. image
In recent years, deep artificial neural networks have attracted much attention and have been applied in various fields because they surpass the parameter fitting effect of traditional methods under the condition of da...
详细信息
In recent years, deep artificial neural networks have attracted much attention and have been applied in various fields because they surpass the parameter fitting effect of traditional methods under the condition of data convergence. On the other hand, limited transmission bandwidth and storage capacity make image compression necessary in communication. Here, a compression algorithm that combines the K-means clustering algorithm with the neural network algorithm is proposed. First, the pixel points of the image are clustered by K-means algorithm in order to reduce the amount of data input to the neural network algorithm. Secondly, neural network is used to extract image features which realizes further compression. The experiment results show that the peak signal-to-noise ratio (PSNR) is 33.48 dB at most with compression ratio at 32:1. The ablation experiment shows that the run time speeds up 9.5% compared to the algorithm without K-means clustering. Comprehensive comparison experiment shows that the average PSNR is 30.09 dB, which is larger than other baseline approaches. The proposed algorithm is an efficient solution for image compression. A compression algorithm combining K-means algorithm with neural network algorithm is proposed. Simulation results show that this proposed algorithm is superior to other baseline algorithms in terms of peak signal-to-noise ratio value and subjective ***
Video color style transfer aims to transform the color style of an original video by using a reference style image. Most existing methods employ neural networks, which come with challenges like opaque transfer process...
详细信息
Video color style transfer aims to transform the color style of an original video by using a reference style image. Most existing methods employ neural networks, which come with challenges like opaque transfer processes and limited user control over the outcomes. Typically, users cannot fine-tune the resulting images or videos. To tackle this issue, we introduce a method that predicts specific parameters for color style transfer using two images. Initially, we train a neural network to learn the corresponding color adjustment parameters. When applying style transfer to a video, we fine-tune the network with key frames from the video and the chosen style image, generating precise transformation parameters. These are then applied to convert the color style of both images and videos. Our experimental results demonstrate that our algorithm surpasses current methods in color style transfer quality. Moreover, each parameter in our method has a specific, interpretable meaning, enabling users to understand the color style transfer process and allowing them to perform manual fine-tuning if desired.
Content-based image retrieval (CBIR) techniques are widely used for extracting specific images from large databases. Recent studies have shown that edge features, alongside colors, align closely with human perception ...
详细信息
Content-based image retrieval (CBIR) techniques are widely used for extracting specific images from large databases. Recent studies have shown that edge features, alongside colors, align closely with human perception in CBIR. However, most CBIR approaches detect edges using linear methods like gradients, which do not align with how the human visual system (HVS) perceives edges. Bioinspired approaches, based on HVS, have proven more effective for edge detection. This study introduces a novel bioinspired spiking neural network (SNN)-based edge detection method for CBIR. The proposed method reduces computational costs by approximately 2.5 times compared to existing SNN models and offers a simpler, easily integrated structure. When integrated into CBIR techniques using conventional edge detection methods (Sobel, Canny, and image derivatives), it increased the mean precision on the Corel-1k dataset by over 3%. These results indicate that the proposed method is effective for edge-based CBIR.
With the development of deep learning techniques, single-image super-resolution methods based on deep learning have made great progress, enabling significant improvements in image quality and detail reproduction. Howe...
详细信息
With the development of deep learning techniques, single-image super-resolution methods based on deep learning have made great progress, enabling significant improvements in image quality and detail reproduction. However, deep convolutional neural networks are often complicated and hard to be understood, and the computational cost limits the application of the models in practical situations. In order to deploy the network on mobile devices with very limited computing power, we build a refined image super-resolution model based on shuffle learning. Based on extensive experimental results on image super-resolution using three widely used datasets, our model not only achieves high scores on the peak signal-to-noise ratio/structural similarity index matrix, but also is simpler and easier to be implemented than other image super-resolution models.
暂无评论