Dam-Al

Hoang, Dinh-Hieu; Diep, Gia‐Han; Tran, Minh-Triet; Le, Ngan

doi:10.1145/3477314.3507112

Cited by 6 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Third, to prevent overfitting and hasten the training of the suggested model, a sequential spectral-spatial attention module has been integrated into a residual block. DAM-AL [25] proposes dilated attention mechanism that has spatial structural features of low level with spatial attention and context features of high level with channel-wise attention. Moreover, dilated attention network has to skip connections and atrous block layers to support capturing both high and low-level features correlating between foreground and background and recapturing the change of context relationship around the object boundary by a prototypical network iteratively.…”

Section: Attention Networkmentioning

confidence: 99%

See 1 more Smart Citation

BADANet: Boundary Aware Dilated Attention Network for Face Parsing

Raghavendra,

Abhilash,

Nookala

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Over the past few years, deep learning techniques have revolutionized the field of face parsing by utilizing massive datasets to generate high-level features and achieve outstanding performance. Usually, these techniques involve Convolutional Neural Networks (CNNs) to derive features from the input image and then a decoder network to forecast the semantic labels for every pixel. Even with complex deep convolutional neural networks (DCNN), incorrect parsing leads to a semantic gap between identical features, especially at boundary levels. Face parsing has encountered additional challenges due to the inclusion of pose, illumination and facial expressions. In order to address these issues, a Boundary Aware Dilated Attention Network (BADANet) is introduced which explores the use of multiscale techniques to improve the accuracy and robustness of the per-pixel frames. BADANet's dilated attention module has been combined with a lightweight backbone to achieve exceptional results on LaPa, CelebAMask-HQ, and iBugMask. Extensive evaluations of the proposed method demonstrate its performance to be on par with various state-of-the-art techniques. The pipeline for the BADANet is available at https://github.com/abhigoku10/BADANet

show abstract

Section: Attention Networkmentioning

confidence: 99%

“…Here Dilated attention mechanism [25] is employed as an attention block to extract multi-scale features. Low-level structural features are extracted by low-level layers(conv1, conv2) of the attention network and high-level context is produced by layers conv3, conv4 and conv5.…”

Section: B Attention Blockmentioning

confidence: 99%