The detection and identification of impact craters on a planetary surface are crucially important for studies autonomous navigation. Crater refers to finding in given image, whereas means actually mapping them particular reference craters. However, no method is available simultaneously detecting identifying with sufficient accuracy robustness. Thus, this study proposes novel end-to-end fully convolutional neural network (CNN), namely, CraterIDNet, which takes remotely sensed images any size...

Since the observables at particular time instants in a temporal sequence exhibit dependencies, they are not independent samples. Thus, it is plausible to apply i.i.d. assumption-based dimensionality reduction methods data. This paper presents novel supervised approach for data, called Linear Sequence Discriminant Analysis (LSDA). It learns linear discriminative projection of feature vectors sequences lower-dimensional subspace by maximizing separability classes such that entire holistically...

Light scattering is a pervasive problem in many areas. Recently, deep learning was implemented speckle reconstruction. To better investigate the key feature extraction and generalization abilities of networks for sparse pattern reconstruction, we develop "one-to-all" self-attention armed convolutional neural network (SACNN). It can extract local global properties different types patterns, unseen glass diffusers, untrained detection positions. We quantitatively analyzed performance ability...

Abstract Given the impressive achievement of convolutional neural networks (CNNs) in grasping image priors from extensive datasets, they have been widely utilized for tasks related to restoration. Recently, there is significant progress another category architectures—Transformers. These models demonstrated remarkable performance natural language and higher‐level vision applications. Despite their ability address some CNNs limitations, such as restricted receptive fields adaptability issues,...

Abstract Lipreading is a visual recognition of speech by using lip movement, which aims to recognise phrases and sentences spoken talking face without the audio. However, existed models for lipreading suffer from slow training speed insufficient performance. To accelerate model lipreading, batch group algorithm proposed, groups all data different frames. In addition, 3D‐MouthNet‐BLSTM‐CTC architecture proposed improve It bases on 3D convolutional neural network, MouthNet, two Bi‐LSTMs, CTC...

Effective fingerprint classification not only can provide an important index mechanism for large database, but also improve the efficiency and performance of AFIS. At present, because traditional Poincare method detection more false singular points weaker anti-noise problem, this paper presents a based on continuously directional image symmetrical axis. Compared with algorithms, algorithm has following two aspects improved: firstly, exhibits good continuity, well gradualness, excellent...

With the blossom of light field cameras, a number studies followed. Depth estimation becomes one most important problems in and numerous algorithms have been proposed to achieve accurate depth map from image. However, few take occlusions into consideration, especially for multi-occlusion. This makes incorrect object boundaries. In this paper, based on theories photo-consistency Lambertian surfaces correspondence between spatial space angular space, we method handle field. The occlusion edges...

Existing algorithms for enhancing low-light images predominantly focus on the region, which leads to over-enhancement of glare and high complexity algorithm makes it difficult apply embedded devices. In this paper, a GS-AGC based regional luminance perception is proposed. The indirect human eye’s vision was taken into account. All similar pixels that satisfied region were extracted, adaptive adjustment processing performed different regions images. proposed method evaluated experimentally...

Local dimming technology focuses on improving the contrast ratio of displayed images for a great visual perception. It consists backlight extraction and pixel compensation. Considering single existing algorithm can hardly adapt to with diverse characteristics rich details, we propose stronger adaptive local method details preservation in this paper. This method, combining advantages some methods introducing combination subjective evaluation objective evaluation, obtains adaptation compared...

The detection of partial discharge (PD) is one the most important methods for cable insulation condition diagnosis. When PD signal detected in cable, it often indicates that there are defects insulation. and pattern classification can provide information about type severity defect, which guide maintenance ensure safe operation cable. In this paper, four kinds artificial defect model were prepared simulation experiments HV including corona discharge, internal suspended surface discharge. Data...

In recent years, with the continuous development of deep learning-based image classification and recognition, artificial intelligence has become most representative keyword our time, intelligent video surveillance technology also a current research hotspot. Intelligent system is mainly developed on computer vision, pattern recognition machine learning techniques. It first captures videos or images by camera, then filters redundant data, finally extract effective features from data for...

Abstract This paper presents a rapid local backlight dimming method for interlaced video to improve the efficiency of displaying video. In proposed method, we combine line average deinterlacing algorithm with algorithm. order reduce computational complexity and power consumption, three novel methods are proposed. The first is add sparse process deinterlaced image. Second, improved dynamic threshold used calculate factor. And third, odd field even same frame use brightness do dimming....

Frequency diverse array (FDA) capable of diversity in both frequency and phase domain can improve range angle resolution, which is attractive Direction Angle (DOA) estimation. But practical realization, DOA estimation performance will be degraded by the mutual coupling effect between antennas. In this paper, a block sparse representation based calibration method proposed for FDA Steering vector parameterization effective utilize inherent structure received signals. And l <sub...

Under the view point of complex networks, a Markov chain can he described by weight directed graph, in which different edges is labeled transfer probabilities ofjumping from current state to future states. The original quantile graph method based on transition probability with equal number existing data points strategy. It ignored amplitude information during constructing associate network. In this brief paper, modified distances interval dividing strategy proposed. Sensitivity analysis two...