Snapshot compressive imaging (SCI) refers to systems where multiple frames are mapped into a single measurement, with video and hyperspectral as two representative applications. Though exciting results of high-speed videos images have been demonstrated, the poor reconstruction quality precludes SCI from wide This paper aims boost via exploiting high-dimensional structure in desired signal. We build joint model integrate nonlocal self-similarity video/hyperspectral rank minimization approach...

In this paper, we present a compositional and dynamic model for face aging. The represents faces in each age group by hierarchical And-or graph, which And nodes decompose into parts to describe details (e.g., hair, wrinkles, etc.) crucial perception Or represent large diversity of alternative selections. Then instance is transverse the graph-parse graph. Face aging modeled as Markov process on parse graph representation. We learn parameters from annotated data set stochasticity dynamics...

Single-pixel imaging (SPI) is a novel technique that captures 2D images using photodiode, instead of conventional array sensors. SPI has high signal-to-noise ratio, wide spectral range, low cost, and robustness to light scattering. Various algorithms have been proposed for reconstruction, including linear correlation methods, the alternating projection (AP) method, compressive sensing (CS) based methods. However, there no comprehensive review discussing respective advantages, which important...

Abstract Existing multispectral imagers mostly use available array sensors to separately measure 2D data slices in a 3D spatial-spectral cube. Thus they suffer from low photon efficiency, limited spectrum range and high cost. To address these issues, we propose conduct imaging using single bucket detector, take full advantage of its sensitivity, wide range, cost, small size light weight. Technically, utilizing the detector’s fast response, scene’s information is multiplexed into dense 1D...

Snapshot compressive imaging (SCI) aims to capture the high-dimensional (usually 3D) images using a 2D sensor (detector) in single snapshot. Though enjoying advantages of low-bandwidth, low-power and low-cost, applying SCI large-scale problems (HD or UHD videos) our daily life is still challenging. The bottleneck lies reconstruction algorithms; they are either too slow (iterative optimization algorithms) not flexible encoding process (deep learning based end-to-end networks). In this paper,...

Modeling the long-term face aging process is of great importance for recognition and animation, but there a lack sufficient sequences model learning. To address this problem, we propose CONcatenational GRaph Evolution (CONGRE) model, which adopts decomposition strategy in both spatial temporal aspects to learn patterns from partially dense databases. In aspect, build graphical representation, human decomposed into mutually interrelated subregions under anatomical guidance. evolution above...

In this paper, we propose a reweighted low-rank matrix recovery method and demonstrate its application for robust image restoration. the literature, principal component pursuit solves problem via convex program of mixed nuclear norm ℓ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> norm. Inspired by minimization sparsity enhancement, reweighting singular values to enhance low rank matrix. An efficient iterative scheme is proposed...

Recently Fourier Ptychography (FP) has attracted great attention, due to its marked effectiveness in leveraging snapshot numbers for spatial resolution large field-of-view imaging. To acquire high signal-to-noise-ratio (SNR) images under angularly varying illuminations subsequent reconstruction, FP requires long exposure time, which largely limits practical applications. In this paper, based on the recently reported Wirtinger flow algorithm, we propose an iterative optimization framework...

Fourier ptychography (FP) is a recently reported technique, for large field-of-view and high-resolution imaging. Specifically, FP captures set of low-resolution images, under angularly varying illuminations, stitches them together in the domain. One FP's main disadvantages its long capturing process, due to requisite number incident illumination angles. In this Letter, utilizing sparsity natural images domain, we propose highly efficient method, termed adaptive (AFP), which applies content...

We consider the reconstruction problem of video snapshot compressive imaging (SCI), which captures high-speed videos using a low-speed 2D sensor (detector). The underlying principle SCI is to modulate sequential frames with different masks and then these encoded are integrated into on thus can be low-speed. On one hand, enjoys advantages low-bandwidth, low-power low-cost. other applying large-scale problems (HD or UHD videos) in our daily life still challenging bottlenecks lies algorithm....

Multi-spectral imaging is a fundamental tool characterizing the constituent energy of scene radiation. However, current multi-spectral video cameras cannot scale up beyond megapixel resolution due to optical constraints and complexity reconstruction algorithms. To circumvent above issues, we propose tens-of-megapixel handheld videography approach (THETA), with proof-of-concept camera achieving 65-megapixel 12 wavebands within visible light range. The high performance brought by multiple...

In this paper we present a dynamic model for simulating face aging process. We adopt high resolution grammatical model[1] and augment it with age hair features. This represents all images by multi-layer And-Or graph integrates three most prominent aspects related to changes: global appearance changes in style shape, deformations effects of facial components, wrinkles at various zones. Then is modeled as Markov process on representation which learned from large dataset. Given an input image,...

Abstract Fourier ptychographic microscopy (FPM) is a novel computational coherent imaging technique for high space-bandwidth product imaging. Mathematically, (FP) reconstruction can be implemented as phase retrieval optimization process, in which we only obtain low resolution intensity images corresponding to the sub-bands of sample’s (HR) spatial spectrum and aim retrieve complex HR spectrum. In real setups, measurements always suffer from various degenerations such Gaussian noise, Poisson...

Specular reflection exists widely in photography and causes the recorded color deviating from its true value, thus, fast high quality highlight removal a single nature image is of great importance. In spite progress past decades removal, achieving wide applicability to large diversity scenes quite challenging. To handle this problem, we propose an analytic solution based on L2 chromaticity definition corresponding dichromatic model. Specifically, paper derives normalized model for pixels...

Abstract Combining spectral imaging with compressive sensing (CS) enables efficient data acquisition by fully utilizing the intrinsic redundancies in natural images. Current multispectral imagers, which are mostly based on array sensors (e.g, CCD or CMOS), suffer from limited range and relatively low photon efficiency. To address these issues, this paper reports a scheme single-pixel detector. Inspired spatial resolution redundancy of current light modulators (SLMs) relative to target...

Signal capture is at the forefront of perceiving and understanding environment; thus, imaging plays a pivotal role in mobile vision. Recent unprecedented progress artificial intelligence (AI) has shown great potential development advanced platforms with new devices. Traditional systems based on "capturing images first processing afterward" mechanism cannot meet this explosive demand. On other hand, computational (CI) are designed to high-dimensional data an encoded manner provide more...

The problem of modeling an animatable 3D human head avatar under lightweight setups is significant importance but has not been well solved. Existing representations either perform in the realism portrait images synthesis or accuracy expression control, both. To address problem, we introduce a novel hybrid explicit-implicit representation, Facial Model Conditioned Neural Radiance Field, which integrates expressiveness NeRF and prior information from parametric template. At core our...

Computational ghost imaging (CGI) achieves single-pixel by using a Spatial Light Modulator (SLM) to generate structured illuminations for spatially resolved information encoding. The speed of CGI is limited the modulation frequency available SLMs, and sets back its practical applications. This paper proposes bypass this limitation trading off SLM's redundant spatial resolution multiplication frequency. Specifically, pair galvanic mirrors sweeping across high SLM multiply within gap between...

Structured illumination technique enhances the lateral resolution by projecting non-uniform intensity patterns on a sample. In typical implementation, three phase shifts (0, 2π/3, 4π/3) are needed for each orientation of sinusoidal pattern, and 3 different orientations to double bandwidth isotopically in Fourier domain. To this end, 9 incoherent images acquisition process. paper, we discuss an imaging strategy structured demonstrate use modified ptychographic procedure reducing number...

Correlation is widely used to reconstruct the object image in ghost imaging (GI). But it only offers a linear proportion of signal-to-noise ratios (SNR) number measurements. We develop Gerchberg-Saxton-like technique for GI reconstruction this manuscript. The proposed takes advantage integral property Fourier transform, and treat captured data as constraints reconstruction. numerically experimentally demonstrate technique, observe nonlinear growth SNR value with respect measurements...

Multimodal microscopes either use multiple cameras or a single camera to multiplex different modes spatially. The former needs expertise demanding alignment and the latter suffers from limited spatial resolution. Here, we report an alignment-free full-resolution simultaneous fluorescence phase imaging approach using single-pixel detectors. By combining reference-free interferometry with scheme, employ structured illumination encode of sample into two detection arms, then conduct...

Abstract The comprehensive analysis of biological specimens brings about the demand for capturing spatial, temporal and spectral dimensions visual information together. However, such high-dimensional video acquisition faces major challenges in developing large data throughput effective multiplexing techniques. Here, we report snapshot hyperspectral volumetric microscopy that computationally reconstructs profiles high-resolution volumes ~1000 μm × 1000 500 at rate by a novel four-dimensional...

Capturing both fine structure and high dynamics of the scene is demanded in many applications. However, such throughput recording requires significant transmission bandwidth large storage. Off-the-shelf super-resolution temporal compressive sensing can partially address this challenge, but directly concatenating two techniques fails to boost because artifact accumulation magnification sequential reconstruction. In Letter, we propose an encoded capturing approach simultaneously increase...

Under low-light environment, handheld photography suffers from severe camera shake under long exposure settings. Although existing deblurring algorithms have shown promising performance on well-exposed blurry images, they still cannot cope with snapshots. Sophisticated noise and saturation regions are two dominating challenges in practical deblurring: the former violates Gaussian or Poisson assumption widely used most thus degrades their badly, while latter introduces non-linearity to...