2020/12/01 by Nithin Raghavan, Raghavan, Nithin, Punarjay Chakravarty +3
Computer Science · Physics and Astronomy · #Advanced Neural Network Applications #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Machine Learning (cs.LG) #Model Reduction and Neural Networks #cs.CV #cs.LG
paper · pdf · doi:10.48550/arxiv.2012.00238
9 pages, 3 figures
arxiv created 2020/12/01 · openalex publication_date 2020/12/01 · arxiv updated 2020/12/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/04
Image-based learning methods for autonomous vehicle perception tasks require large quantities of labelled, real data in order to properly train without overfitting, which can often be incredibly costly. While leveraging the power of simulated data can potentially aid in mitigating these costs, networks trained in the simulation domain usually fail to perform adequately when applied to images in the real domain. Recent advances in domain adaptation have indicated that a shared latent space assumption can help to bridge the gap between the simulation and real domains, allowing the transference of the predictive capabilities of a network from the simulation domain to the real domain. We demonstrate that a twin VAE-based architecture with a shared latent space and auxiliary decoders is able to bridge the sim2real gap without requiring any paired, ground-truth data in the real domain. Using only paired, ground-truth data in the simulation domain, this architecture has the potential to generate perception tasks such as depth and segmentation maps. We compare this method to networks trained in a supervised manner to indicate the merit of these results.