Self-supervised learning for autonomous vehicles perception: A\n conciliation between analytical and learning methods

2019/10/03 by Florent Chiaroni, Chiaroni, Florent, Mohamed-Cherif Rahal +5 · 1 citation
Computer Science · Engineering · #Advanced Neural Network Applications #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #Image and Video Processing (eess.IV) #Machine Learning (cs.LG) #Robotics (cs.RO) #Robotics and Sensor-Based Localization #Video Surveillance and Tracking Methods #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.1910.01636

openalex publication_date 2019/10/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Nowadays, supervised deep learning techniques yield the best state-of-the-art\nprediction performances for a wide variety of computer vision tasks. However,\nsuch supervised techniques generally require a large amount of manually labeled\ntraining data. In the context of autonomous vehicles perception, this\nrequirement is critical, as the distribution of sensor data can continuously\nchange and include several unexpected variations. It turns out that a category\nof learning techniques, referred to as self-supervised learning (SSL), consists\nof replacing the manual labeling effort by an automatic labeling process.\nThanks to their ability to learn on the application time and in varying\nenvironments, state-of-the-art SSL techniques provide a valid alternative to\nsupervised learning for a variety of different tasks, including long-range\ntraversable area segmentation, moving obstacle instance segmentation, long-term\nmoving obstacle tracking, or depth map prediction. In this tutorial-style\narticle, we present an overview and a general formalization of the concept of\nself-supervised learning (SSL) for autonomous vehicles perception. This\nformalization provides helpful guidelines for developing novel frameworks based\non generic SSL principles. Moreover, it enables to point out significant\nchallenges in the design of future SSL systems.\n

Citations

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• The expectation-maximization algorithm
• Random sample consensus
• Distinctive Image Features from Scale-Invariant Keypoints
• Depth Map Prediction from a Single Image using a Multi-Scale Deep Network
• Estimating the class prior and posterior from noisy positives and unlabeled data
• Autonomous driving challenge: To Infer the property of a dynamic object\n based on its motion pattern using recurrent neural network
• 1 year, 1000 km: The Oxford RobotCar dataset
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Self-Supervised Learning for Stereo Matching with Self-Improving Ability
• Dimensionality-Driven Learning with Noisy Labels
• Deep Clustering for Unsupervised Learning of Visual Features
• Self-Supervised Visual Feature Learning With Deep Neural Networks: A Survey
• Bayesian SegNet: Model Uncertainty in Deep Convolutional Encoder-Decoder Architectures for Scene Understanding

Cited by

• Deep Learning Perspective of Scene Understanding in Autonomous Robots

Related

• Persistent self-supervised learning principle: from stereo to monocular vision for obstacle avoidance
• Recent Advancements in Deep Learning Applications and Methods for Autonomous Navigation: A Comprehensive Review
• Learning Long-Range Perception Using Self-Supervision from Short-Range\n Sensors and Odometry
• Motion Inspired Unsupervised Perception and Prediction in Autonomous Driving
• Self-supervised classification of dynamic obstacles using the temporal information provided by videos