Filters








8 Hits in 4.9 sec

FC^4: Fully Convolutional Color Constancy with Confidence-Weighted Pooling

Yuanming Hu, Baoyuan Wang, Stephen Lin
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
To overcome this problem, we present a fully convolutional network architecture in which patches throughout an image can carry different confidence weights according to the value they provide for color  ...  Improvements in color constancy have arisen from the use of convolutional neural networks (CNNs).  ...  To address this problem, we propose a fully convolutional network, called FC 4 , where the patches in an input image can differ in influence over the color constancy estimation.  ... 
doi:10.1109/cvpr.2017.43 dblp:conf/cvpr/HuWL17 fatcat:67m4wc5azngtrhpgky7pggbx74

Monte Carlo Dropout Ensembles for Robust Illumination Estimation [article]

Firas Laakom, Jenni Raitoharju, Alexandros Iosifidis, Jarno Nikkanen, Moncef Gabbouj
2020 arXiv   pre-print
Computational color constancy is a preprocessing step used in many camera systems.  ...  The main aim is to discount the effect of the illumination on the colors in the scene and restore the original colors of the objects.  ...  To overcome this limitation, recently FC 4 [17] and BoCF [19] were proposed. FC 4 [17] method relies on a fully convolutional topology using a confidence-weighted pooling layer.  ... 
arXiv:2007.10114v1 fatcat:q72rpryjpnclxecx27jzpptdme

Artificial Color Constancy via GoogLeNet with Angular Loss Function [article]

Oleksii Sidorov
2019 arXiv   pre-print
Color Constancy is the ability of the human visual system to perceive colors unchanged independently of the illumination.  ...  FC 4 is a fully-convolutional network that allows using images without resizing or cropping; also, it uses confidence-weighed pooling which helps to avoid ambiguity through assignment to each patch confidence  ...  weights according to the value they provide for color constancy estimation.  ... 
arXiv:1811.08456v2 fatcat:4tjotbdhhndpfp277rw7evww5y

INTEL-TAU: A Color Constancy Dataset [article]

Firas Laakom, Jenni Raitoharju, Alexandros Iosifidis, Jarno Nikkanen, Moncef Gabbouj
2020 arXiv   pre-print
Furthermore, this paper benchmarks several color constancy approaches on the proposed dataset.  ...  Furthermore, the effect of color shading for mobile images can be evaluated with INTEL-TAU dataset, as both corrected and uncorrected versions of the raw data are provided.  ...  Furthermore, we evaluated the performance of the following learning-based methods: Fast Fourier Color Constancy (FFCC) [7] , Fully Convolutional Color Constancy With Confidence-Weighted Pooling (FC 4  ... 
arXiv:1910.10404v5 fatcat:b4atpqnyjvhfjhcr262tvhr2dq

INTEL-TAU: A Color Constancy Dataset

Firas Laakom, Jenni Raitoharju, Jarno Nikkanen, Alexandros Iosifidis, Moncef Gabbouj
2021 IEEE Access  
Furthermore, this paper benchmarks several color constancy approaches on the proposed dataset.  ...  INDEX TERMS Color constancy, dataset, illumination estimation, regression. 39560 This work is licensed under a Creative Commons Attribution 4.0 License.  ...  Furthermore, we evaluated the performance of the following learning-based methods: Fast Fourier Color Constancy (FFCC) [7] , Fully Convolutional Color Constancy With Confidence-Weighted Pooling (FC 4  ... 
doi:10.1109/access.2021.3064382 fatcat:nlxssjjvirdf5ja4fdfaiso3mq

A Benchmark for Temporal Color Constancy [article]

Yanlin Qian and Jani Käpylä and Joni-Kristian Kämäräinen and Samu Koskinen and Jiri Matas
2020 arXiv   pre-print
Temporal Color Constancy (CC) is a recently proposed approach that challenges the conventional single-frame color constancy.  ...  In temporal CC, multiple frames from the view finder sequence are used to estimate the color. However, there are no realistic large scale temporal color constancy datasets for method evaluation.  ...  TCC-Net baseline (Model G) is a fully 2D convolutional architecture that is the best found architecture for illuminant estimation in temporal color constancy.  ... 
arXiv:2003.03763v1 fatcat:kkzqs6w4jjh4hkxkrwfe5vxwfy

Robust channel-wise illumination estimation [article]

Firas Laakom, Jenni Raitoharju, Jarno Nikkanen, Alexandros Iosifidis, Moncef Gabbouj
2021 arXiv   pre-print
In this work, we propose a novel color constancy uncertainty estimation approach that augments the trained model with an auxiliary branch which learns to predict the error based on the feature representation  ...  Recently, Convolutional Neural Networks (CNNs) have been widely used to solve the illuminant estimation problem and have often led to state-of-the-art results.  ...  For the supervised approaches, we report the results of Fast Fourier Color Constancy (FFCC) [9] and the following five CNN-based approaches: Fully Convolutional Color Constancy (FC 4 ) [13] , Bianco  ... 
arXiv:2111.05681v1 fatcat:a25jvidsmfb3rlawt2ex4u3ms4

Image color correction, enhancement, and editing [article]

Mahmoud Afifi
2021 arXiv   pre-print
In particular, we propose auto image recoloring methods to generate different realistic versions of the same camera-rendered image with new colors.  ...  This thesis presents methods and approaches to image color correction, color enhancement, and color editing.  ...  We could not find learning-based models trained on sRGB images for illuminant estimation except for the fully convolutional color constancy with confidence-weighted pooling (FC4) model [177] .  ... 
arXiv:2107.13117v1 fatcat:uf6qv7iux5hqxmgxxb4olvvtzy