Lateral refinement network for contour detection

Lin, Chia-Wen; Cui, Linhao; Li, Fuzhang; Cao, Yijun

doi:10.1016/j.neucom.2020.06.069

Cited by 19 publications

(17 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The structure of the US layer is shown in Figure 6C . Compared with the current popular methods ( Liu et al, 2017 ; Cao et al, 2020 ; He et al, 2020 ; Lin et al, 2020 ), we do not use the deconvolution for up-sampling but choose the sub-pixel convolution method to build the US layer. Considering the down-sampling of patch merging, which increases the number of channels and decreases the resolution, we use the sub-pixel convolution for up-sampling.…”

Section: Methodsmentioning

confidence: 99%

“…Recently, an excellent edge detection method ( Wang et al, 2018 ; Cao et al, 2020 ; Lin et al, 2020 ) obtained richer feature information by fusing the multi-scale features extracted by the backbone. LRC ( Lin et al, 2020 ) shows that hierarchical fusion through a fusion module is beneficial for achieving more abundant features. In this study, the design of a decoding network refers to the hierarchical fusion method.…”

Section: Methodsmentioning

confidence: 99%

“…In the LRC network with a more complex decoding network, the SPN has greatly improved the three evaluation indicators. Experiment 3: To explore the validity of the feature fusion module (FFM), we used the refinement block proposed in LRC (Lin et al, 2020) to replace our FFM. The experimental results are summarized in Table 3.…”

Section: Ablation Experimentsmentioning

confidence: 99%

“…Cao et al ( Cao et al, 2020 ) proposed the deep refinement network (DRC) by designing a refinement module to build a decoding network to achieve the fusion of features with different scales. Lin et al ( Lin et al, 2020 ) proposed a new edge detection method (lateral reinforcement network for contour detection—LRC). They used the decoding module to design a deeper decoding network to fuse multi-scale features.…”

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

DPED: Bio-inspired dual-pathway network for edge detection

Chen

Lin

Qiao

2022

Front. Bioeng. Biotechnol.

View full text Add to dashboard Cite

Edge detection is significant as the basis of high-level visual tasks. Most encoder-decoder edge detection methods used convolutional neural networks, such as VGG16 or Resnet, as the encoding network. Studies on designing decoding networks have achieved good results. Swin Transformer (Swin) has recently attracted much attention in various visual tasks as a possible alternative to convolutional neural networks. Physiological studies have shown that there are two visual pathways that converge in the visual cortex in the biological vision system, and that complex information transmission and communication is widespread. Inspired by the research on Swin and the biological vision pathway, we have designed a two-pathway encoding network. The first pathway network is the fine-tuned Swin; the second pathway network mainly comprises deep separable convolution. To simulate attention transmission and feature fusion between the first and second pathway networks, we have designed a second-pathway attention module and a pathways fusion module. Our proposed method outperforms the CNN-based SOTA method BDCN on BSDS500 datasets. Moreover, our proposed method and the Transformer-based SOTA method EDTER have their own performance advantages. In terms of FLOPs and FPS, our method has more benefits than EDTER.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Ablation Experimentsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

DPED: Bio-inspired dual-pathway network for edge detection

Chen

Lin

Qiao

2022

Front. Bioeng. Biotechnol.

View full text Add to dashboard Cite

show abstract

“…The experiment results show that ODS is 0.808 in the BSDS500 dataset. In 2018, Lin et al [74] proposed a lateral refinement network algorithm, which is based on CNN and refinement module, using VGG19 network obtained ODS is 0.816 and 0.761 in the BSDS500 and NYUD-V2 date sets, moreover, using ResNet network obtained ODS is 0.820 and 0.760 in BSDS500 and NYUD-V2 date sets. In 2016, Liu et al [75] designed a fully convolutional neural network based on the structural characteristics of each layer, which can let ODS is 0.806.…”

Section: Edge Detection Technology Based On Cross-layer Multiscale Fusionmentioning

confidence: 99%

A Review of the Edge Detection Technology

Hou¹,

Jia²,

Wang³

et al. 2021

STAIQC

View full text Add to dashboard Cite

The edge detection-based has profoundly inspired recent works in image classification, object detection, segmentation, et al. With the growth of computer vision, the performance of edge detection has been notably improved. In this paper, we concentrate on presenting some edge detection technologies and grouping them into two major categories: classical edge detection technology and deep learning-based edge detection technology. For every team in the classification, the fundamental thoughts are first described, and then we probe beneath into the mainly related literature in recent years. Furthermore, the evaluation indicators are entirely elaborated. In addition, this conclusion is given by a comprehensive summary which provides suggestions on future perspectives.

show abstract