Image Colorization: A Survey and Dataset
Image colorization is the process of estimating RGB colors for grayscale images or video frames to improve their aesthetic and perceptual quality. Deep learning techniques for image colorization have progressed notably over the last decade, calling the need for a systematic survey and benchmarking of these techniques. This article presents a comprehensive survey of recent state-of-the-art deep learning-based image colorization techniques, describing their fundamental block architectures, inputs, optimizers, loss functions, training protocols, and training data \textit{etc.} It categorizes the existing colorization techniques into seven classes and discusses important factors governing their performance, such as benchmark datasets and evaluation metrics. We highlight the limitations of existing datasets and introduce a new dataset specific to colorization. Using the existing datasets and our new one, we perform an extensive experimental evaluation of existing image colorization methods. Finally, we discuss the limitations of existing methods and recommend possible solutions as well as future research directions for this rapidly evolving topic of deep image colorization. Dataset and codes for evaluation are publicly available at https://github.com/saeed-anwar/ColorSurvey
PDF Abstract
Datasets
Introduced in the Paper:
NCD
Used in the Paper:
CIFAR-10
ImageNet
MS COCO
CIFAR-100
Oxford 102 Flower
Places205
COCO-Stuff
ImageNet ctest10k
