rakeshpavan333 / colorization

Colorization Performance Analysis of different models and existing papers

Colorization of B/W images using Deep learning

Abstract

• Given a grayscale (black and white) image as input, we shall attack the problem of hallucinating a plausible color version of the photograph.The problem we want to solve is a fully automatic approach to colorization devoid of any human interference. This project will also compare performances of different models and existing papers .

• Project Proposal: https://github.com/Rakeshpavan333/colorization/blob/master/Colorization.pdf

Highlights

• Each pixel is assigned a color from the appropriate region using a neighborhood matching metric, combined with spatial filtering for improved spatial coherence .

• Pixels with a sufficiently high confidence level are provided as “micro-scribbles” to the optimization-based colorization algorithm of Levin et al. [LLW04 ] .

• Performance analysis over (a neural network model,a CNN model , auto encoder model and an encoder-decoder classifier) .

• For a demo go to : https://github.com/Rakeshpavan333/colorization/tree/master/demo