An Empirical Evaluation of Deep Learning on Highway Driving

Abstract

Numerous groups have applied a variety of deep learning techniques to computer vision problems in highway perception scenarios. In this paper, we presented a number of empirical evaluations of recent deep learning advances. Computer vision, combined with deep learning, has the potential to bring about a relatively inexpensive, robust solution to autonomous driving. To prepare deep learning for industry uptake and practical applications, neural networks will require large data sets that represent all possible driving environments and scenarios. We collect a large data set of highway data and apply deep learning and computer vision algorithms to problems such as car and lane detection. We show how existing convolutional neural networks (CNNs) can be used to perform lane and vehicle detection while running at frame rates required for a real-time system. Our results lend credence to the hypothesis that deep learning holds promise for autonomous driving.

VIL-100: A New Dataset and A Baseline Model for Video Instance Lane Detection

“…Early lane detection methods mostly relied on hand-crafted features, such as color [41,45], edge [9,22,27] and texture [23]. Recently, the use of deep neural networks [39,16,30] has significantly boosted the lane detection performance. In VPGNet [20], vanishing points were employed to guide a multi-task network training for lane detection.…”

End-to-End Deep Learning of Lane Detection and Path Prediction for Real-Time Autonomous Driving

“…Lane detection is an essential task of advanced driver assistance systems (ADAS) in modern vehicles and autonomous driving systems in self-driving cars for lane keeping assistance, departure warning, and centering [1], [2], [3], [4], [5], [6], [7]. There are two major paradigms, namely, classic computer vision and deep learning, for lane detection [1].…”

“…Lane detection is an essential task of advanced driver assistance systems (ADAS) in modern vehicles and autonomous driving systems in self-driving cars for lane keeping assistance, departure warning, and centering [1], [2], [3], [4], [5], [6], [7]. There are two major paradigms, namely, classic computer vision and deep learning, for lane detection [1]. The classic computer vision paradigm requires intensive feature engineering, road modeling, and special case handling and is thus not robust enough to deal with seemingly infinite driving situations, environments, and unexpected obstacles [1].…”

“…There are two major paradigms, namely, classic computer vision and deep learning, for lane detection [1]. The classic computer vision paradigm requires intensive feature engineering, road modeling, and special case handling and is thus not robust enough to deal with seemingly infinite driving situations, environments, and unexpected obstacles [1]. The deep learning paradigm has recently shown considerable progress to alleviate these difficulties [1], [2], [3], [4], [5], [6], [7].…”

Relative stability toward diffeomorphisms indicates performance in deep nets

“…Deep learning algorithms LeCun et al (2015) are now remarkably successful at a wide range of tasks Amodei et al (2016); Huval et al (2015); Mnih et al (2013); Shi et al (2016); Silver et al (2017). Yet, understanding how they can classify data in large dimensions remains a challenge.…”