Open Access Open Access  Restricted Access Subscription Access

An Approach to Convert Grayscale Images to Color

Urvashi Kar, Prajval B, Shaikh Arif Mohd Mondal, Chandan D S, Dr. T H Sreenivas


This paper investigates the coloring problems that occur while transforming a grayscale image into a color image. Earlier works on colorization techniques involve approaches to choose colors from a palette of RGB and transferring them on to the gray image. In other methods it either requires a lot of color scribbling on the black and white image or a huge dataset of color reference for the image in particular to the era it belongs to. We use convolutional neural networks along with a feature extractor and the Inception-ResNet-v2 pre-trained classifier model for higher efficiency in coloring. Our neural network is combined with the classifier that increases the performance of similar images. We train our neural network on images from Unsplash, an image collection website, that are available as a public dataset

Full Text:



C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna. “Rethinking the inception architecture for computer

vision” arXiv preprint arXiv:1512.00567, 2015.

D. Blasi, R. Recupero, “Fast colorization of gray images,” Eurographics Italian Chapter, 2003.

T. Welsh, M. Ashikhmin, and K. Mueller. “Transferring Color to Greyscale Images,” in Proceedings of ACM SIGGRAPH, San Antonio, 2002, pp 227-280.

J. Hwang and Y. Zhou, “Image colorization with deep convolutional neural networks,” Stanford University, Tech. Rep.,2016.

Z. Cheng, Q. Yang, B. Sheng, “Deep colorization,” in Proceedings of the International Conference on Computer Vision, 2015.


  • There are currently no refbacks.