Attention transfer from human to neural networks for road object detection in winter

Abstract

As an essential feature of autonomous road vehicles, obstacle detection must be executed on a real‐time onboard platform with high accuracy. Cameras are still the most commonly used sensors in autonomous driving. Most detections using cameras are based on convolutional neural networks. In this regard, a recent teacher–student approach, called transfer learning, has been used to improve the neural network training process. This approach has only been used with a neural network acting as a teacher to the best of our knowledge. This paper proposes a novel way of improving training data based on attention transfer by getting the attention map from a human. The proposed method allows the dataset size reduction by 50%, which leads to up to a 60% decline in the training time. The experimental results indicate that the proposed method can enhance the F1‐score of the network by up to 10% in winter conditions.

How to deal with glare for improved perception of Autonomous Vehicles

“…Additionally, we have compared the performance of the proposed method on a real dataset [31] captured from an autonomous car using a GoPro HERO5 camera. The testing dataset in this paper consists of a subset of images from [31] that contain strong sources of glare i.e all the nighttime images, made publicly available 3 . We have compared the proposed method with a variety of techniques applicable to the glare reduction problem.…”