An automatic image colorization system using deep convolutional auto encoders compared with CNN algorithm

The primary goal of this research is to achieve maximum accuracy using a deep convolutional autoencoder architecture in comparison with the CNN algorithm. An automatic image colorization system was implemented using autoencoders and Convolutional Neural Networks, both of which were evaluated on a sample size of 5. The iteration value was determined based on G-power calculations, targeting a power value of 0.8 with a 95% confidence interval. The Autoencoder method is documented to possess an accuracy of 94.11%, which is higher than the accuracy of the CNN approach, as it is observed to be 93.40%. The two algorithms exhibit a statistically significant difference, as indicated by a 2-tailed p-value of 0.037 (p