Face completion with Hybrid Dilated Convolution

Fang, Yuchun; Li, Yifan; Tu, Xiaokang; Tan, Taifeng; Wang, Xin

doi:10.1016/j.image.2019.115664

Cited by 36 publications

(19 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this section, we demonstrate the performance and generated results of our model through experiments. Due to the work of this paper is mainly focused on generator, in order to make comparative experiments more meaningful, we mainly use different generators (CNN [6], Dilated CNN [19], Partial CNN [20], Hybrid dilated CNN [21] and Laplacian pyramid CNN [22] ) to compare with our model.…”

Section: Methodsmentioning

confidence: 99%

“…In order to evaluate the performance of our model, we compare our network with CE network [6], GL network [20], PConv [19] and HDC [21]. The CE uses convolution layers for feature extracting, GL uses dilated convolution layers for feature extracting, PConv uses partial convolution for feature extracting, HDC uses U-Net based convolution for feature extracting, LPC uses Laplacian pyramid based convolution and our network uses pixel LSTM layers for feature extracting.…”

Section: Comparison With State-of-the-arts Methodsmentioning

confidence: 99%

“…For example, the convolutional neural network (CNN) [6], dilated convolution [19] and partial convolution [20] enhanced the performance of the completion network by improving the performance of encoder responsible for feature extraction in the generator. Furthermore, in order to generate more detailed textures, hybrid dilated convolution (HDC) [21] and laplacian-pyramid-based convolution (LPC) [22] used U-Net based method and multi-scale theory to extract texture features in facial images. However, in the existing methods, the edges of the completion region regularly have structural and semantic discontinuities, since the spatial dependence of images is not used to extract features.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Facial Image Completion Using Bi-Directional Pixel LSTM

Xiang

Zhang

2020

IEEE Access

View full text Add to dashboard Cite

Structural features of facial images directly affect the performance of the image completion model. However, most existing work does not make full use of spatial dependence to extract features, and cause the semantics and structure of completion being inconsistent with the context. This paper addresses this issue using a bi-directional pixel long-short time memory (LSTM) network. Specifically, it consists of two LSTM subnetworks and can simultaneously scan the input image row by row or column by column, thereby the extracted features contain the dependence information among rows or among columns. Through a fusion operation of these features, the complete spatial dependence information is included. In addition, the parameters of the decoder and discriminator are automatically adjusted to accommodate the proposed bi-directional pixel LSTM. Finally, compared with existing state-of-the-art methods, experimental results show our superior performance.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Comparison With State-of-the-arts Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Facial Image Completion Using Bi-Directional Pixel LSTM

Xiang

Zhang

2020

IEEE Access

View full text Add to dashboard Cite

show abstract

“…2) Newly proposed approaches only stake the DConv layers [8] with same dilation rates linearly. This will cause the gridding effect [9] and redundant feature computation for multi-scale objects [10], both in natural images and aerial images. 3) Most existing approaches ignore the scene information of background.…”

Section: Introductionmentioning

confidence: 99%

Lightweight Dual-Task Networks For Crowd Counting In Aerial Images

Tian

Duan

Zhang

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

As a research hotspot of computer vision, crowd counting methods have achieved success in natural images. But crowd counting in aerial images are rarely explored, and existing methods do not perform well because of the higher resolution, smaller object scale and more complex scene. Therefore, this paper proposes a lightweight dual-task network (LDNet) for crowd counting, which only uses bifurcated structure to overcome these new challenges in aerial images without complicated pipelines. To realize this, a complete but efficient Guidance Branch is proposed to assist Counting Branch in fitting crowd distribution. Furthermore, a scene attention mechanism is used to consider the complex scene information, which are never considered by existing methods. Our LD-Net outperforms existing methods on aerial crowd counting dataset (Visdrone), and gets better or comparable results on natural crowd counting datasets (UCF CC 50, UCF QNRF, ShanghaiTech Part A).

show abstract

“…Liu et al [ 22 ] proposed a liver CT sequence image segmentation algorithm GIU-Net, which combines an improved U-Net neural network model with graph cutting. The U-Net-based method proposed by Fang et al [ 23 ] combines hybrid dilated convolution (HDC) and spectral normalization, which can use sharp structures and fine textures to fill missing areas of any shape. Hong et al [ 24 ] used U-Net to develop a novel segmentation framework suitable for deep WMH.…”

Section: Introductionmentioning

confidence: 99%

An Ore Image Segmentation Method Based on RDU-Net Model

Xiao

Liu

et al. 2020

Sensors

View full text Add to dashboard Cite

The ore fragment size on the conveyor belt of concentrators is not only the main index to verify the crushing process, but also affects the production efficiency, operation cost and even production safety of the mine. In order to get the size of ore fragments on the conveyor belt, the image segmentation method is a convenient and fast choice. However, due to the influence of dust, light and uneven color and texture, the traditional ore image segmentation methods are prone to oversegmentation and undersegmentation. In order to solve these problems, this paper proposes an ore image segmentation model called RDU-Net (R: residual connection; DU: DUNet), which combines the residual structure of convolutional neural network with DUNet model, greatly improving the accuracy of image segmentation. RDU-Net can adaptively adjust the receptive field according to the size and shape of different ore fragments, capture the ore edge of different shape and size, and realize the accurate segmentation of ore image. The experimental results show that compared with other U-Net and DUNet, the RDU-Net has significantly improved segmentation accuracy, and has better generalization ability, which can fully meet the requirements of ore fragment size detection in the concentrator.

show abstract

Face completion with Hybrid Dilated Convolution

Cited by 36 publications

References 17 publications

Facial Image Completion Using Bi-Directional Pixel LSTM

Facial Image Completion Using Bi-Directional Pixel LSTM

Lightweight Dual-Task Networks For Crowd Counting In Aerial Images

An Ore Image Segmentation Method Based on RDU-Net Model

Contact Info

Product

Resources

About