ICNet for Real-Time Semantic Segmentation on High-Resolution Images

Abstract

We focus on the challenging task of real-time semantic segmentation in this paper. It finds many practical applications and yet is with fundamental difficulty of reducing a large portion of computation for pixel-wise label inference. We propose an image cascade network (ICNet) that incorporates multi-resolution branches under proper label guidance to address this challenge. We provide in-depth analysis of our framework and introduce the cascade feature fusion unit to quickly achieve high-quality segmentation. Our system yields real-time inference on a single GPU card with decent quality results evaluated on challenging datasets like Cityscapes, CamVid and COCO-Stuff.

ECCV 2018

References in corpus (5)

• Pruning Filters for Efficient ConvNets
• Learning Deconvolution Network for Semantic Segmentation
• YOLO9000: Better, Faster, Stronger
• DSD: Dense-Sparse-Dense Training for Deep Neural Networks
• COCO-Stuff: Thing and Stuff Classes in Context

Cited by in corpus (25)

• ContextNet: Exploring Context and Detail for Semantic Segmentation in Real-time
• BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation
• PointSeg: Real-Time Semantic Segmentation Based on 3D LiDAR Point Cloud
• Improving Fast Segmentation With Teacher-student Learning
• DFANet: Deep Feature Aggregation for Real-Time Semantic Segmentation
• Gated CRF Loss for Weakly Supervised Semantic Image Segmentation
• Real-time Semantic Image Segmentation via Spatial Sparsity
• Guided Upsampling Network for Real-Time Semantic Segmentation
• AI Benchmark: Running Deep Neural Networks on Android Smartphones
• LEDNet: A Lightweight Encoder-Decoder Network for Real-Time Semantic Segmentation
• Knowledge Adaptation for Efficient Semantic Segmentation
• SceneAdapt: Scene-based domain adaptation for semantic segmentation using adversarial learning
• Boundary-Aware Network for Fast and High-Accuracy Portrait Segmentation
• Efficient Segmentation: Learning Downsampling Near Semantic Boundaries
• DeLS-3D: Deep Localization and Segmentation with a 3D Semantic Map
• Towards Multi-class Object Detection in Unconstrained Remote Sensing Imagery
• ESNet: An Efficient Symmetric Network for Real-time Semantic Segmentation
• ShelfNet for Fast Semantic Segmentation
• Dynamic Video Segmentation Network
• Context-aware Padding for Semantic Segmentation
• EADNet: Efficient Asymmetric Dilated Network for Semantic Segmentation
• Efficient Semantic Segmentation using Gradual Grouping
• GANtruth - an unpaired image-to-image translation method for driving scenarios
• ESFNet: Efficient Network for Building Extraction from High-Resolution Aerial Images
• Compact retail shelf segmentation for mobile deployment