Abstract

In this paper, automatic black and white image colorization method has been proposed. The study is based on the best-known deep learning algorithm CNN (Convolutional neural network). The Model that developed taking the input in gray scale and predict the color of image based on the dataset that trained on it. The color space used in this work is Lab Color space the model takes the L channel as the input and the ab channels as the output. The Image Net dataset used and random selected image have been used to construct a mini dataset of images that contains 39,604 images splitted into 80% training and 20% testing. The proposed method has been tested and evaluated on samples images with Mean-squared error and peak signal to noise ratio and reached an average of MSE= 51.36 and PSNR= 31.

References

[1] Levin, A., D. Lischinski, and Y. Weiss. Colorization using optimization. in ACMtransactions on graphics (tog). 2004. ACM.
[2] Yatziv, L. and G.J.I.t.o.i.p. Sapiro, Fast image and video colorization using chrominanceblending. 2006. 15(5): p. 1120-1129.
[3] Horiuchi, T.J.I. and V. Computing, Colorization algorithm using probabilistic relaxation.2004. 22(3): p. 197-202.
[4] Takahama, T., T. Horiuchi, and H. Kotera. Improvement on colorization accuracy bypartitioning algorithm in cielab color space. in Pacific-Rim Conference on Multimedia.2004. Springer.
[5] LeCun, Y., Y.J.T.h.o.b.t. Bengio, and n. networks, Convolutional networks for images,speech, and time series. 1995. 3361(10): p. 1995.
[6] LeCun, Y.J.U.h.y.l.c.e.l., LeNet-5, convolutional neural networks. 2015. 20: p. 5.
[7] Hiley, L., A. Preece, and Y.J.a.p.a. Hicks, Explainable Deep Learning for VideoRecognition Tasks: A Framework & Recommendations. 2019.
[8] Hubel, D.H. and T.N.J.T.J.o.p. Wiesel, Receptive fields and functional architecture ofmonkey striate cortex. 1968. 195(1): p. 215-243.
[9] Fukushima, K.J.B.c., Neocognitron: A self-organizing neural network model for amechanism of pattern recognition unaffected by shift in position. 1980. 36(4): p. 193-202.
[10] Matsugu, M., et al., Subject independent facial expression recognition with robust facedetection using a convolutional neural network. 2003. 16(5-6): p. 555-559.
[11] Deng, J., et al. Imagenet: A large-scale hierarchical image database. in 2009 IEEEconference on computer vision and pattern recognition. 2009. Ieee.