Photorealistic Synthetic Data Generation For AI-Based Feature Development

AI feature development for automated driving applications is heavily reliant on large quantities of diverse data. Generative Adversarial Networks (GANs) are now widely used for photo-realistic image synthesis. However, in applications where a simulated image needs to be translated into a realistic image (sim-to-real), GANs trained on unpaired data from the two domains are susceptible to failure in semantic content retention as the image is translated from one domain to the other. This failure mode is more pronounced in cases where the real data lacks content diversity, resulting in a content mismatch between the two domains - a situation often encountered in real-world deployment. This presentation will discuss the role of the discriminator's receptive field in GANs for unsupervised image-to-image translation with mismatched data, and study its effect on semantic content retention. The presentation will also show how targeted synthetic data augmentation - combining the benefits of gaming engine simulations and sim-to-real GANs - can help fill gaps in static datasets for vision tasks such as parking slot detection, lane detection and monocular depth estimation. Prior knowledge in computer vision and deep learning will be helpful in getting the most out of this session.