Abstract:In recent years, intelligent driving navigation and security monitoring have made considerable progress with the help of deep Convolutional Neural Networks (CNNs). As one of the state-of-the-art perception approaches, semantic segmentation unifies distinct detection tasks widely desired by both autonomous driving and security monitoring. Currently, semantic segmentation shows remarkable efficiency and reliability in standard scenarios such as daytime scenes with favorable illumination conditions. However, in face of adverse conditions such as the nighttime, semantic segmentation loses its accuracy significantly. One of the main causes of the problem is the lack of sufficient annotated segmentation datasets of nighttime scenes. In this paper, we propose a framework to alleviate the accuracy decline when semantic segmentation is taken to adverse conditions by using Generative Adversarial Networks (GANs). To bridge the daytime and nighttime image domains, we made key observation that compared to datasets in adverse conditions, there are considerable amount of segmentation datasets in standard conditions such as BDD and our collected ZJU datasets. Our GAN-based nighttime semantic segmentation framework includes two methods. In the first method, GANs were used to translate nighttime images to the daytime, thus semantic segmentation can be performed using robust models already trained on daytime datasets. In another method, we use GANs to translate different ratio of daytime images in the dataset to the nighttime but still with their labels. In this sense, synthetic nighttime segmentation datasets can be generated to yield models prepared to operate at nighttime conditions robustly. In our experiment, the later method significantly boosts the performance at the nighttime evidenced by quantitative results using Intersection over Union (IoU) and Pixel Accuracy (Acc). We show that the performance varies with respect to the proportion of synthetic nighttime images in the dataset, where the sweet spot corresponds to most robust performance across the day and night. The proposed framework not only makes contribution to the optimization of visual perception in intelligent vehicles, but also can be applied to diverse navigational assistance systems.

Natural Scene Digit Classification Using Convolutional Neural Networks

Binocular Depth Estimation Using Convolutional Neural Network With Siamese Branches.

See Clearer at Night: Towards Robust Nighttime Semantic Segmentation Through Day-Night Image Conversion

Scene Classification Of High Resolution Remote Sensing Images Using Convolutional Neural Networks

An Efficient and Lightweight Convolutional Neural Network for Remote Sensing Image Scene Classification

Scene Classification in Indoor Environments for Robots using Context Based Word Embeddings

Urban Remote Sensing Scene Recognition Based on Lightweight Convolution Neural Network

Natural Scene Recognition Based on Superpixels and Deep Boltzmann Machines

Understand Scene Categories by Objects: A Semantic Regularized Scene Classifier Using Convolutional Neural Networks.

Scene classification with improved AlexNet model

Multi-Scale Depthwise Separable Convolution for Semantic Segmentation in Street–Road Scenes

A Convolutional Recurrent Neural-Network-Based Machine Learning for Scene Text Recognition Application

An End-to-End Trainable Multi-Column CNN for Scene Recognition in Extremely Changing Environment

DeepScene: Scene classification via convolutional neural network with spatial pyramid pooling

Automatic Urban Scene-Level Binary Change Detection Based on a Novel Sample Selection Approach and Advanced Triplet Neural Network

Compressed Holistic Convolutional Neural Network-based Descriptors for Scene Recognition

Convolutional Neural Networks Applied to House Numbers Digit Classification

Traffic-Sign Detection and Classification in the Wild

Transferring Deep Convolutional Neural Networks for the Scene Classification of High-Resolution Remote Sensing Imagery

Sentiment Prediction in Scene Images Via Convolutional Neural Networks

A Combined Deep Learning Model for the Scene Classification of High-Resolution Remote Sensing Image