Lidar-Camera Co-Training for Semi-Supervised Road Detection

Abstract

Recent advances in the field of machine learning and computer vision have enabled the development of fast and accurate road detectors. Commonly such systems are trained within a supervised learning paradigm where both an input sensor's data and the corresponding ground truth label must be provided. The task of generating labels is commonly carried out by human annotators and it is notoriously time consuming and expensive. In this work, it is shown that a semi-supervised approach known as co-training can provide significant F1-score average improvements compared to supervised learning. In co-training, two classifiers acting on different views of the data cooperatively improve each other's performance by leveraging unlabeled examples. Depending on the amount of labeled data used, the improvements ranged from 1.12 to 6.10 percentage points for a camera-based road detector and from 1.04 to 8.14 percentage points for a lidar-based road detector. Lastly, the co-training algorithm is validated on the KITTI road benchmark, achieving high performance using only 36 labeled training examples together with several thousands unlabeled ones.

Cross-Modality Domain Adaptation for Freespace Detection: A Simple yet Effective Baseline

“…We use Virtual KITTI as source dataset and the original training set of KITTI-Road as target dataset. It is shown that our method outperforms all the UDA methods and one of the semi-supervised methods RGB36-super [4]. Besides, our method achieves close AP and MaxF to the fully supervised methods, further indicating the large potential of our method.…”

“…To overcome these issues, some methods introduce other modalities of data as supplement of information, such as Lidar [4,11,23] and depth [6,16,60,61]. Gu et al [23] progressively adapts Lidar data space into RGB data space to effectively fuse features of the two modalities.…”