DeepAI AI Chat
Log In Sign Up

DFANet: Deep Feature Aggregation for Real-Time Semantic Segmentation

by   Hanchao Li, et al.
Megvii Technology Limited

This paper introduces an extremely efficient CNN architecture named DFANet for semantic segmentation under resource constraints. Our proposed network starts from a single lightweight backbone and aggregates discriminative features through sub-network and sub-stage cascade respectively. Based on the multi-scale feature propagation, DFANet substantially reduces the number of parameters, but still obtains sufficient receptive field and enhances the model learning ability, which strikes a balance between the speed and segmentation performance. Experiments on Cityscapes and CamVid datasets demonstrate the superior performance of DFANet with 8× less FLOPs and 2× faster than the existing state-of-the-art real-time semantic segmentation methods while providing comparable accuracy. Specifically, it achieves 70.3% Mean IOU on the Cityscapes test dataset with only 1.7 GFLOPs and a speed of 160 FPS on one NVIDIA Titan X card, and 71.3% Mean IOU with 3.4 GFLOPs while inferring on a higher resolution image.


page 4

page 7


BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation

Semantic segmentation requires both rich spatial information and sizeabl...

Not All Pixels Are Equal: Difficulty-aware Semantic Segmentation via Deep Layer Cascade

We propose a novel deep layer cascade (LC) method to improve the accurac...

Feature Pyramid Encoding Network for Real-time Semantic Segmentation

Although current deep learning methods have achieved impressive results ...

Waterfall Atrous Spatial Pooling Architecture for Efficient Semantic Segmentation

We propose a new efficient architecture for semantic segmentation, based...

Revisiting Multi-Scale Feature Fusion for Semantic Segmentation

It is commonly believed that high internal resolution combined with expe...

DRBANET: A Lightweight Dual-Resolution Network for Semantic Segmentation with Boundary Auxiliary

Due to the powerful ability to encode image details and semantics, many ...

Bi-direction Context Propagation Network for Real-time Semantic Segmentation

Spatial details and context correlations are two types of critical infor...