Multi-modality image fusion provides more comprehensive and sophisticated information in modern medical diagnosis, remote sensing, video surveillance, so on. This paper presents a novel multi-modality method based on phase congruency local Laplacian energy. In the proposed method, non-subsampled contourlet transform is performed pairs to decompose source images into high-pass low-pass subbands. The subbands are integrated by congruency-based rule that can enhance detailed features of fused...

Infrared small target detection (IRSTD) has recently benefitted greatly from U-shaped neural models. However, largely overlooking effective global information modeling, existing techniques struggle when the high similarities with background. We present a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">S</i> patial-channel xmlns:xlink="http://www.w3.org/1999/xlink">C</i> ross xmlns:xlink="http://www.w3.org/1999/xlink">T</i> ransformer...

Convolutional neural networks have recently been used for multi-focus image fusion. However, due to the lack of labeled data supervised training such networks, existing methods resorted adding Gaussian blur in focused images simulate defocus and generate synthetic with ground-truth learning. Moreover, they classify pixels as or defocused leverage results construct fusion weight maps which then necessitates a series post-processing steps. In this paper, we present unsupervised end-to-end...

Infrared and visible image fusion technique is a popular topic in analysis because it can integrate complementary information obtain reliable accurate description of scenes. Multiscale transform theory as signal representation method widely used fusion. In this paper, novel infrared proposed based on spectral graph wavelet (SGWT) bilateral filter. The main novelty study that SGWT for On the one hand, source images are decomposed by its domain. approach not only effectively preserves details...

Detecting key frames in videos is a common problem many applications such as video classification, action recognition and summarization. These tasks can be performed more efficiently using only handful of rather than the full video. Existing frame detection approaches are mostly designed for supervised learning require manual labelling large corpus training data to train models. Labelling requires human annotators from different backgrounds annotate which not expensive time consuming but...

The aim of multi-focus image fusion technology is to integrate different partially focused images into one all-focused image. To realize this goal, a new method based on guided filter proposed and an efficient salient feature extraction presented in paper. Furthermore, primarily the main objective present work. Based extraction, first used acquire smoothing containing most sharpness regions. obtain initial map, we compose mixed focus measure by combining variance intensities energy gradient...

Fusion of infrared and visible images is a significant research area in image analysis computer vision. The purpose fusion to combine the complementary information source into fused image. Thus, it vital efficiently represent important choose rational rules. To achieve this aim, an method using multiscale directional nonlocal means (MDNLM) filter proposed paper. MDNLM combines feature preserving edge by with capacity capturing bank, which can effectively intrinsic geometric structure images....

Hepatocyte transplantation is considered to be an alternative orthotopic liver transplantation. Cells can used bridge patients waiting for a donor organ, decrease mortality in acute failure, and support metabolic diseases. The limited availability of primary human hepatocytes such applications has led the generation hepatocyte-like cells from various adult stem or precursor cells. aim this study was generate adipose-derived mesenchymal (Ad-MSCs) clinical applications, which are available...

Infrared small moving target detection is one of the crucial techniques in infrared search and tracking systems. This paper presents a novel method for image sequence with complicated background. The key points are given as follows: (1) since mainly depends on incoherence between background, proposed separate from background according to morphological feature diversity background; (2) considering continuity motion time domain, trajectory extracted by RX filter random projection. experiments...

Thermal infrared (TIR) imaging has been an indispensable tool in surveillance and remote sensing fields due to the characteristic of this spectrum that enables system detect relatively warm targets, especially low-light conditions. However, acquired TIR images often suffer from observable stripe noise, which reduces target detectability some extent. To remove noise keep image details, letter proposes a novel method combines spectral processing technology with image-guidance mechanism....

Convolutional neural networks have recently been used for multi-focus image fusion. However, some existing methods resorted to adding Gaussian blur focused images, simulate defocus, thereby generating data (with ground-truth) supervised learning. Moreover, they classify pixels as 'focused' or 'defocused', and use the classified results construct fusion weight maps. This then necessitates a series of post-processing steps. In this paper, we present an end-to-end learning approach directly...

Hyperspectral imaging (HSI) has been widely investigated within the context of computational due to high dimensional challenges for direct imaging. However, existing HSI approaches are mostly designed visible near-infrared waveband, whereas less attention paid mid-infrared spectral range. In this paper, we report a novel compressive system extend application domain digital micromirror device (MIR-DMD). our system, modified MIR-DMD is combined with an off-the-shelf infrared spectroradiometer...

Detecting representative frames in videos based on human actions is quite challenging because of the combined factors pose action and background. This paper addresses this problem formulates key frame detection as one finding video that optimally maximally contribute to differentiating underlying category from all other categories. To end, we introduce a deep two-stream ConvNet for learns directly predict location frames. Our idea automatically generate labeled data CNN learning using...

The presence of stripe nonuniformity severely degrades the image quality and affects performance in many infrared (IR) sensing applications. Prior works correct by using similar spatial representations, which inevitably damage some detailed structures image. In this paper, we instead take advantage spectral prior noise to solve its correction problem single IR We first analyse significant difference between stripes utilize knowledge characterize nonuniformity. Then a two-stage filtering...

Stripe noise is a common and unwelcome pattern in various thermal infrared (TIR) image data including conventional TIR images remote sensing spectral images. Most existing stripe removal (destriping) methods are often difficult to keep good robust efficacy dealing with the real-life complex cases. In this paper, based on intrinsic properties of noise, we propose novel two-stage transform domain destriping method called Fourier anomaly detection fusion (ADSF). Considering principal...

Orbital angular momentum (OAM) entangled photons propagating through non-Kolmogorov turbulence are studied by numerical simulations. Here, the paper uses multiphase screen model, especially focusing on influences of azimuthal mode and parameters (i.e., generalized exponent, outer scale turbulence, inner turbulence) entanglement evolution in weak scintillation regime. The results indicate that mode, have obvious OAM entanglement. However, influence can be ignored.

Hyperspectral images (HSIs) often suffer from degradation caused by mixed noise, leading to a decline in the performance of subsequent advanced applications. To eliminate noise and improve image quality, transformer-based approaches have been successfully employed. Nevertheless, these strategies involve large-scale modeling tedious layer normalization, which causes inefficiencies during denoising process. Additionally, neglect local spectral correlations HSIs damages physical properties...