PSANet: Point-wise Spatial Attention Network for Scene Parsing

Zhao, Hengshuang; Zhang, Yi; Liu, Shu; Shi, Jianping; Loy, Chen Change; Lin, Dahua; Jia, Jiaya

doi:10.1007/978-3-030-01240-3_17

Cited by 1,005 publications

(597 citation statements)

References 34 publications

Supporting

Mentioning

547

Contrasting

Order By: Relevance

“…3.2, APNB is much more efficient than a standard non-local block. We hereby give a quantitative efficiency comparison between our APNB and a generic non-local block in the following aspects: GFLOPs, GPU memory (MB) and GPU computation time (ms Besides the comparison of the single block efficiency, we also provide the whole network efficiency comparison with the two most advanced methods, PSANet [48] and DenseA-SPP [36], in terms of inference time (s), GPU occupation with batch size set to 1 (MB) and the number of parameters (Million). According to Tab.…”

Section: Efficiency Comparison With Non-local Blockmentioning

confidence: 99%

Asymmetric Non-Local Neural Networks for Semantic Segmentation

Zhu

Mengdu

Bai

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

671

362

View full text Add to dashboard Cite

The non-local module works as a particularly useful technique for semantic segmentation while criticized for its prohibitive computation and GPU memory occupation. In this paper, we present Asymmetric Non-local Neural Network to semantic segmentation, which has two prominent components: Asymmetric Pyramid Non-local Block (APNB) and Asymmetric Fusion Non-local Block (AFNB). APNB leverages a pyramid sampling module into the nonlocal block to largely reduce the computation and memory consumption without sacrificing the performance. AFNB is adapted from APNB to fuse the features of different levels under a sufficient consideration of long range dependencies and thus considerably improves the performance. Extensive experiments on semantic segmentation benchmarks demonstrate the effectiveness and efficiency of our work. In particular, we report the state-of-the-art performance of 81.3 mIoU on the Cityscapes test set. For a 256 × 128 input, APNB is around 6 times faster than a non-local block on GPU while 28 times smaller in GPU running memory occupation. Code is available at: https://github.com/ MendelXu/ANN.git. * Equal contribution † Corresponding author < l a t e x i t s h a 1 _ b a s e 6 4 = " h Z k z g Y O / g V + m e P P y R w 8 d X I Q M H C U = " > A A A B 6 3 i c b Z D L S g M x F I Z P 6 q 3 W W 9 W l m 2 A R X E i Z c a P L g h u X F e w F 2 q F k 0 k w n N M k M S U Y o Q 1 / B j Q t F 3 P p C 7 n w b M + 0 s t P W H

show abstract

Section: Efficiency Comparison With Non-local Blockmentioning

confidence: 99%

Asymmetric Non-Local Neural Networks for Semantic Segmentation

Zhu

Mengdu

Bai

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

671

362

View full text Add to dashboard Cite

show abstract

“…The results and comparison are illustrated in Table 3. ACFNet, which uses only train-fine data, outperforms previous work PSANet [48] for about 2.2% and even better than most methods that also employ the validation set for training. While using both train-fine and val-fine data for training, ACFNet outperforms the previous methods [41,48,43,42] for a large margin and achieves new stateof-the-art of 81.85% mIoU.…”

Section: Comparing With the State-of-the-artmentioning

confidence: 86%

“…Besides, our special thanks go to Yuchen Sun, Xueyu Song, Ru Zhang, Yuhui Yuan and the anonymous reviewers for the discussion and their helpful advice. road sidewalk building wall fence pole traffic light traffic sign vegetation terrain sky person rider car truck bus train motorcycle bicycle PSPNet †[47] 78.4 -------------------PSANet † [48] 78 .6 ------------------ [42] 78.9 -…”

Section: Acknowledgmentmentioning

confidence: 99%

ACFNet: Attentional Class Feature Network for Semantic Segmentation

Zhang

Chen

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

275

128

View full text Add to dashboard Cite

Recent works have made great progress in semantic segmentation by exploiting richer context, most of which are designed from a spatial perspective. In contrast to previous works, we present the concept of class center which extracts the global context from a categorical perspective. This class-level context describes the overall representation of each class in an image. We further propose a novel module, named Attentional Class Feature (ACF) module, to calculate and adaptively combine different class centers according to each pixel. Based on the ACF module, we introduce a coarse-to-fine segmentation network, called Attentional Class Feature Network (ACFNet), which can be composed of an ACF module and any off-the-shell segmentation network (base network). In this paper, we use two types of base networks to evaluate the effectiveness of ACFNet. We achieve new state-of-the-art performance of 81.85% mIoU on Cityscapes dataset with only finely annotated data used for training.

show abstract

“…Deeplab-v2 [9] 70.4 RefineNet-Res101 [43] 73.6 DSSPN-Universal [41] 76.6 GCN [56] 76.9 DepthSet [35] 78.2 PSPNet [75] 78.4 AAF [34] 79.1 DFN [72] 79.3 PSANet [76] 80.1 DenseASPP-DenseNet161 [71] 80.6…”

Section: Methods Mean Ioumentioning

confidence: 99%

Semantic Correlation Promoted Shape-Variant Context for Segmentation

Ding

Jiang

Shuai

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

173

View full text Add to dashboard Cite

Context is essential for semantic segmentation. Due to the diverse shapes of objects and their complex layout in various scene images, the spatial scales and shapes of contexts for different objects have very large variation. It is thus ineffective or inefficient to aggregate various context information from a predefined fixed region. In this work, we propose to generate a scale-and shape-variant semantic mask for each pixel to confine its contextual region. To this end, we first propose a novel paired convolution to infer the semantic correlation of the pair and based on that to generate a shape mask. Using the inferred spatial scope of the contextual region, we propose a shape-variant convolution, of which the receptive field is controlled by the shape mask that varies with the appearance of input. In this way, the proposed network aggregates the context information of a pixel from its semantic-correlated region instead of a predefined fixed region. Furthermore, this work also proposes a labeling denoising model to reduce wrong predictions caused by the noisy low-level features. Without bells and whistles, the proposed segmentation network achieves new state-of-the-arts consistently on the six public segmentation datasets.

show abstract

PSANet: Point-wise Spatial Attention Network for Scene Parsing

Cited by 1,005 publications

References 34 publications

Asymmetric Non-Local Neural Networks for Semantic Segmentation

Asymmetric Non-Local Neural Networks for Semantic Segmentation

ACFNet: Attentional Class Feature Network for Semantic Segmentation

Semantic Correlation Promoted Shape-Variant Context for Segmentation

Contact Info

Product

Resources

About