Large-scale labeled datasets are generally necessary for successfully training a deep neural network in the computer vision domain. In order to avoid the costly and tedious work of manually annotating image datasets, self-supervised learning methods have been proposed to learn general visual features automatically. In this paper, we first focus on image colorization with generative adversarial networks (GANs) because of their ability to generate the most realistic colorization results. Then, via transfer learning, we use this as a proxy task for visual understanding. Particularly, we propose to use conditional GANs (cGANs) for image colorization and transfer the gained knowledge to two other downstream tasks, namely, multilabel image classification and semantic segmentation. This is the first time that GANs have been used for self-supervised feature learning through image colorization. Through extensive experiments with the COCO and Pascal datasets, we show an increase of 5% for the classification task and 2.5% for the segmentation task. This demonstrates that image colorization with conditional GANs can boost other downstream tasks’ performance without the need for manual annotation.
Automatic image colorization is an image-to-image translation problem where a model is able to automatically convert a one-dimensional grayscale image to a three-dimensional color image. Inspiration is taken from the Pix2Pix image-to-image translation model and then a pre-trained MobileNet model is integrated into the generator of the Conditional Generative Adversarial Network (cGAN). The goal is to use transfer learning for the purpose of achieving satisfactory results with less training data while at the same time decreasing training time and improving performance. Moreover, quantitative and qualitative analysis is performed. The results show that by using transfer learning the modified model achieves a PSNR score of 24.21 and a SSIM score of 0.9374 and with that improves upon the baseline.