FBSNet: A Fast Bilateral Symmetrical Network for Real-Time Semantic Segmentation

Gao, Guangwei; Xu, Guoan; Li, Juncheng; Yu, Yi; Lu, Huimin; Yang, Jian

doi:10.1109/tmm.2022.3157995

Cited by 36 publications

(16 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Global Semantic Module: Inspired by the context module of BiSeNet, we also utilize projection convolution and average pooling to construct our module for capturing more global context information [55]. As Fig.…”

Section: Ablation For Sifmentioning

confidence: 99%

DuFNet: Dual Flow Network of Real-Time Semantic Segmentation for Unmanned Driving Application of Internet of Things

Duan¹,

Liu²,

Li³

et al. 2023

Computer Modeling in Engineering &Amp; Sciences

View full text Add to dashboard Cite

The application of unmanned driving in the Internet of Things is one of the concrete manifestations of the application of artificial intelligence technology. Image semantic segmentation can help the unmanned driving system by achieving road accessibility analysis. Semantic segmentation is also a challenging technology for image understanding and scene parsing. We focused on the challenging task of real-time semantic segmentation in this paper. In this paper, we proposed a novel fast architecture for real-time semantic segmentation named DuFNet. Starting from the existing work of Bilateral Segmentation Network (BiSeNet), DuFNet proposes a novel Semantic Information Flow (SIF) structure for context information and a novel Fringe Information Flow (FIF) structure for spatial information. We also proposed two kinds of SIF with cascaded and paralleled structures, respectively. The SIF encodes the input stage by stage in the ResNet18 backbone and provides context information for the feature fusion module. Features from previous stages usually contain rich low-level details but high-level semantics for later stages. The multiple convolutions embed in Parallel SIF aggregate the corresponding features among different stages and generate a powerful global context representation with less computational cost. The FIF consists of a pooling layer and an upsampling operator followed by projection convolution layer. The concise component provides more spatial details for the network. Compared with BiSeNet, our work achieved faster speed and comparable performance with 72.34% mIoU accuracy and 78 FPS on Cityscapes Dataset based on the ResNet18 backbone.

show abstract

Section: Ablation For Sifmentioning

confidence: 99%

DuFNet: Dual Flow Network of Real-Time Semantic Segmentation for Unmanned Driving Application of Internet of Things

Duan¹,

Liu²,

Li³

et al. 2023

Computer Modeling in Engineering &Amp; Sciences

View full text Add to dashboard Cite

show abstract

“…Similar to ENet [18] and FBSNet [2], we first use three 3×3 convolution to extract the early-stage features and then we employ the same downsampling policy with ENet initial block. The downsampling operation can be seen in Figure 6.…”

Section: Downsampling Operationmentioning

confidence: 99%

“…At the end of this part, visual comparison results will be presented. The performance of our EARMNet is tested with several state-of-art works in this part on the Cityscapes datasets: SegNet [6], Enet [18], SQNet [28], ESPNet [19], CGNet [38], ContextNet [39], EDANet [40], Fast-SCNN [42], Fast-SCNN [42], BiseNet [1], ICNet [43], DABNet [36], LEDNet [10], FBSNet [2], DFANet [44], FDDWNet [45] and MSCFNet [21]. We can learn from Table 5 and Table 6, the comparison results show that our EARMNet achieves a good balance between prediction accuracy and efficiency.…”

Section: Performance Evaluation On Cityscapesmentioning

confidence: 99%

See 1 more Smart Citation

EARMNet:A Fast and Accurate Semantic Segmentation Network with Lightweight Efficient Asymmetric Residual Module

Shen

Zhai

Chen³

et al. 2023

Preprint

View full text Add to dashboard Cite

Recently years, semantic segmentation has had extensive usage in many scenarios. Most real-world applications both require high prediction accuracy and running efficiency. To this end, we present a novel lightweight semantic segmentation method to pursue a balance between prediction accuracy and running efficiency. We first design our Efficient Asymmetric Residual Module (EARM) to capture local and context features with a lower computation cost. Following that, we introduce a novel Image Decomposition Attention Mechanism (IDAM) to fuse the feature map from different stages. Based on EARM, we arrange two EARM blocks with gradually increasing dilation rates as the main part of our method. Besides, we show a Lightweight Mutual-Attention Decoder (LMAD) to recover the detailed information effectively. Extensive experiments on Cityscapes and Camvid datasets show that our model achieves 73.3\% and 71.7\% mIoU along with 150 and 131 FPS on the Cityscapes and Camvid datasets respectively

show abstract

“…This dual‐branch down‐sampling structure has been adopted in many works because of its effective balance of feature quality and inference speed. FBSNet [35] optimizes the dual‐branch down‐sampling structure. It designs a lightweight bottleneck residual unit and stacks this unit to compose its semantic branches.…”

Section: Related Workmentioning

confidence: 99%

SA‐BiSeNet: Swap attention bilateral segmentation network for real‐time inland waterways segmentation

Zhang

Bao

2022

IET Image Processing

View full text Add to dashboard Cite

The technology for autonomous navigation on inland waterways is worth investigating, and navigable water surface segmentation is a key part of this technology. Semantic segmentation methods based on deep learning are able to distinguish between water surface areas and non-water surface areas. However, existing semantic segmentation methods cannot meet the requirements of the water surface segmentation task in terms of both segmentation precision and real-time performance. In this study, a Swap Attention Bilateral Segmentation Network (SA-BiSeNet) is proposed to improve segmentation performance while ensuring model inference speed by better fusing the two features of the dual-branch down-sampling network using the attention mechanism. Specifically, an innovative Swap Attention Module is designed to model the dependency between the features of the spatial detail branch and the features of the semantic branches, thus expanding the receptive fields of the spatial detail and semantic branches to each other's global contexts. This design can effectively fuse features and thus enhance feature representation. Experiments were conducted on the inland waterway dataset USVInland to verify the performance of SA-BiSeNet in terms of segmentation precision and inference speed, and SA-BiSeNet achieved 93.65% Mean IoU and maintained the same level of fps as the baseline.

show abstract

FBSNet: A Fast Bilateral Symmetrical Network for Real-Time Semantic Segmentation

Cited by 36 publications

References 47 publications

DuFNet: Dual Flow Network of Real-Time Semantic Segmentation for Unmanned Driving Application of Internet of Things

DuFNet: Dual Flow Network of Real-Time Semantic Segmentation for Unmanned Driving Application of Internet of Things

EARMNet:A Fast and Accurate Semantic Segmentation Network with Lightweight Efficient Asymmetric Residual Module

SA‐BiSeNet: Swap attention bilateral segmentation network for real‐time inland waterways segmentation

Contact Info

Product

Resources

About