Automatic Example-Based Image Colorization Using Location-Aware Cross-Scale Matching

Li, Bo; Lai, Yu‐Kun; John, Matthew P.; Rosin, Paul L.

doi:10.1109/tip.2019.2912291

Cited by 48 publications

(26 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Over the last decade, many approaches have been proposed to address colorization issues. Previous work falls into three broad categories, which are scribble-based [8,11,22,24,31,35,38], transfer-based [2,5,14,23,27,37], and fully automatic colorization [19].…”

Section: Related Workmentioning

confidence: 99%

“…Thus, the colorized results were deterministic, and relied heavily on the collection of reference images, which is referred to as a forward problem. We call this kind of method transfer-based [2,5,14,23,27,37]. In addition, researchers proposed scribble-based colorization [8,11,22,24,31,35,38], which works by following users' guidance with manual intervention.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Unpaired medical image colorization using generative adversarial network

Liang

Lee

et al. 2021

Multimed Tools Appl

View full text Add to dashboard Cite

We consider medical image transformation problems where a grayscale image is transformed into a color image. The colorized medical image should have the same features as the input image because extra synthesized features can increase the possibility of diagnostic errors. In this paper, to secure colorized medical images and improve the quality of synthesized images, as well as to leverage unpaired training image data, a colorization network is proposed based on the cycle generative adversarial network (CycleGAN) model, combining a perceptual loss function and a total variation (TV) loss function. Visual comparisons and experimental indicators from the NRMSE, PSNR, and SSIM metrics are used to evaluate the performance of the proposed method. The experimental results show that GAN-based style conversion can be applied to colorization of medical images. As well, the introduction of perceptual loss and TV loss can improve the quality of images produced as a result of colorization better than the result generated by only using the CycleGAN model.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Unpaired medical image colorization using generative adversarial network

Liang

Lee

et al. 2021

Multimed Tools Appl

View full text Add to dashboard Cite

show abstract

“…However, the final results contained numerous color bleeding effects. In order to reduce color bleeding, various improvements have been studied [4], [6]- [19]. For example, salient contours were introduced in [6] to cope with color bleeding artifacts caused by weak object boundaries.…”

Section: Related Workmentioning

confidence: 99%

“…In addition, each vertex is processed equally without taking into account some important geometric properties, such as concavity. On the other hand, the quadratic optimization will result in obvious color bleeding around boundaries due to its low-pass characteristic [4]. Although a feature line field technique is introduced to detect the boundaries, the mechanism of quadratic diffusion still leads to obvious color bleeding around the detected boundaries ( Fig.…”

Section: Introductionmentioning

confidence: 99%

Sparse Graph Regularized Mesh Color Edit Propagation

Ли

Lai

Rosin

2020

IEEE Trans. on Image Process.

Self Cite

View full text Add to dashboard Cite

Mesh color edit propagation aims to propagate the color from a few color strokes to the whole mesh, which is useful for mesh colorization, color enhancement and color editing, etc. Compared with image edit propagation, luminance information is not available for 3D mesh data, so the color edit propagation is more difficult on 3D meshes than images, with far less research carried out. This paper proposes a novel solution based on sparse graph regularization. Firstly, a few color strokes are interactively drawn by the user, and then the color will be propagated to the whole mesh by minimizing a sparse graph regularized nonlinear energy function. The proposed method effectively measures geometric similarity over shapes by using a set of complementary multiscale feature descriptors, and effectively controls color bleeding via a sparse 1 optimization rather than quadratic minimization used in existing work. The proposed framework can be applied for the task of interactive mesh colorization, mesh color enhancement and mesh color editing. Extensive qualitative and quantitative experiments show that the proposed method outperforms the state-of-the-art methods.

show abstract

“…Since just low-level surface highlights are utilized, surface coordinating. Sorour Mohajerani et al, (2019) [7] proposed CPAdv-Net that is motivated by U-Net. CPAdv-Net has a semisymmetrical engineering comprising of two principle arms (ways): the contracting and the extending.…”

Section: Previous Workmentioning

confidence: 99%

Forgery Detection Based on KNN Classifier using SURF Feature Extraction

DHIVYA¹,

SUDHAKAR²

2019

IJRTE

View full text Add to dashboard Cite

Copy move forgery is a standout amongst the most widely recognized for controlling unique pictures. Edge detection is a standout amongst the most contemplated issues in PC vision, yet it remains a difficult undertaking. It is troublesome since frequently the choice for an edge can't be made absolutely dependent on low dimension signs, for example, slope, rather we have to connect all dimensions of data, low, center, and high, so as to choose where to put edges. In this paper we propose a novel adjusted K-derives gathering calculation for edge and thing limit affirmation which we suggest as Speeded-Up Robust Features (SURF).A choice of an edge point is made autonomously at every area in the picture; an extremely huge opening is utilized giving critical setting to every choice. In the coordinating stage, the calculation chooses and consolidates an enormous number of highlights crosswise over various scales so as to gain proficiency with a discriminative model utilizing an all-inclusive rendition of the Putatively Matched Points (Including Outliers) calculation.The coordinating based structure is exceptionally versatile and there are no parameters to tune. The proposed work is pertinent to applications for imitation discovery in various explicit picture areas just as on common pictures. The outcomes are reproduced through the MatlabR2014b programming.

show abstract

Automatic Example-Based Image Colorization Using Location-Aware Cross-Scale Matching

Cited by 48 publications

References 53 publications

Unpaired medical image colorization using generative adversarial network

Unpaired medical image colorization using generative adversarial network

Sparse Graph Regularized Mesh Color Edit Propagation

Forgery Detection Based on KNN Classifier using SURF Feature Extraction

Contact Info

Product

Resources

About