Training and Synthesis Based Vehicle Color Conver sion Technique Using GAN
Aroona Ayub, HyungWon Kim · Journal of Korea Multimedia Society · 2023
While recent Convolutional Neural Networks (CNNs) for object detection have been substantially improved, they require a large amount of annotated data to further improve their accuracy to the level of human. Such annotated data is scarce. Manual annotation of object labels is a time consuming and expensive process. Recently, generative models are being employed to automate the manual annotation and produce diverse training data. This leads to an increase in the accuracy of the target model. This paper presents a method to train a Generative Adversarial Network (GAN) network to translate the vehicle colors in the given dataset. It generates augmented image data by translating the selected vehicle objects to 7 color domains with the least compromise in the quality of generated images. We demonstrate the result of training the proposed GAN model and performance of vehicle color translation using Comprehensive Cars (CompCars) dataset.