A survey of the state-of-the-art in patch-based synthesis

Barnes, Connelly; Zhang, Fang-Lue

doi:10.1007/s41095-016-0064-2

Cited by 61 publications

(26 citation statements)

References 85 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A comprehensive overview of patch‐based synthesis techniques is provided by Barnes et al Now, we get the error map R ERR between R HR and f CNN ( R LR ) from the first step. In this step, we use the patch‐based method based on the error map (denoted as R ERR ) at the view of R HR .…”

Section: Proposed Methodsmentioning

confidence: 99%

A novel refocusing distance measurement method using super‐resolved light field images

Zhao

Hao

2018

Computer Animation & Virtual

View full text Add to dashboard Cite

This paper proposes a novel distance measurement method using superresolved light-field images. We first superresolve the light field using the hybrid cross-resolution input based on PatchMatch and convolutional neural network method, which combines and takes advantage of these two methods. In this way, we can explore the similarity between images and deal with challenging scenes effectively. The scaling factor is up to eight times, which is much larger than the ordinary superresolution scaling factor, and our superresolution method can also maintain a satisfactory accuracy. With the traditional idea of focus ranging, the light-field 3D measurement method is established. A proper definition evaluation function suitable for light-field data is selected to evaluate the imaging quality, and the triangle barycenter interpolation method is proposed to measure 3D scenes. Experimental results demonstrate that the proposed method not only improves the quality of the reconstructed high-resolution light field but also has the ability of the distance measurement. It indicates that the light-field imaging is a promising 3D measurement technique.

show abstract

Section: Proposed Methodsmentioning

confidence: 99%

A novel refocusing distance measurement method using super‐resolved light field images

Zhao

Hao

2018

Computer Animation & Virtual

View full text Add to dashboard Cite

show abstract

“…Image/Video Inpainting. The problem of image inpainting/completion [3] has been studied extensively, with classical approaches focusing on patch-based non-parametric optimization [2,11,12,14,15,23,26,28,39,46] as well as more recent work using deep generative neural net-works [18,31,35,47,48]. On the other hand, the video inpainting problem has received far less attention from research community.…”

Section: Related Workmentioning

confidence: 99%

An Internal Learning Approach to Video Inpainting

Zhang

Mai

Jin

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

View full text Add to dashboard Cite

We propose a novel video inpainting algorithm that simultaneously hallucinates missing appearance and motion (optical flow) information, building upon the recent 'Deep Image Prior' (DIP) that exploits convolutional network architectures to enforce plausible texture in static images. In extending DIP to video we make two important contributions. First, we show that coherent video inpainting is possible without a priori training. We take a generative approach to inpainting based on internal (within-video) learning without reliance upon an external corpus of visual data to train a one-size-fits-all model for the large space of general videos. Second, we show that such a framework can jointly generate both appearance and flow, whilst exploiting these complementary modalities to ensure mutual consistency. We show that leveraging appearance statistics specific to each video achieves visually plausible results whilst handling the challenging problem of long-term consistency.

show abstract

“…Although patch-based appearance transfer methods were successful in synthesizing temporally coherent textures for flow animations, these are limited to 2D setups. For a broad discussion of patch-based texture synthesis works we refer to [Barnes and Zhang 2017].…”

Section: Related Workmentioning

confidence: 99%

Transport-based neural style transfer for smoke simulations

et al. 2019

View full text Add to dashboard Cite

Artistically controlling fluids has always been a challenging task. Optimization techniques rely on approximating simulation states towards target velocity or density field configurations, which are often handcrafted by artists to indirectly control smoke dynamics. Patch synthesis techniques transfer image textures or simulation features to a target flow field. However, these are either limited to adding structural patterns or augmenting coarse flows with turbulent structures, and hence cannot capture the full spectrum of different styles and semantically complex structures. In this paper, we propose the first Transport-based Neural Style Transfer (TNST) algorithm for volumetric smoke data. Our method is able to transfer features from natural images to smoke simulations, enabling general content-aware manipulations ranging from simple patterns to intricate motifs. The proposed algorithm is physically inspired, since it computes the density transport from a source input smoke to a desired target configuration. Our transport-based approach allows direct control over the divergence of the stylization velocity field by optimizing incompressible and irrotational potentials that transport smoke towards stylization. Temporal consistency is ensured by transporting and aligning subsequent stylized velocities, and 3D reconstructions are computed by seamlessly merging stylizations from different camera viewpoints.

show abstract

A survey of the state-of-the-art in patch-based synthesis

Cited by 61 publications

References 85 publications

A novel refocusing distance measurement method using super‐resolved light field images

A novel refocusing distance measurement method using super‐resolved light field images

An Internal Learning Approach to Video Inpainting

Transport-based neural style transfer for smoke simulations

Contact Info

Product

Resources

About