An efficient mesh-based face beautifier on mobile devices

Fan, Xin; Chai, Zhi; Feng, Yuyao; Wang, Yi; Wang, Shengfa; Luo, Zhongxuan

doi:10.1016/j.neucom.2014.09.093

Cited by 5 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our method MCL outperforms most of the state-of-the-art methods, especially on AFLW dataset where a relative error reduction of 3.93% is achieved compared to RecNet. Cascaded CNN estimates the location of each 2 The result is acquired by running the code at https://github.com/seetaface/SeetaFaceEngine/tree/master/FaceAlignment. landmark separately in the second and third level, and every two networks are used to detect one landmark.…”

Section: B Comparison With State-of-the-art Methodsmentioning

confidence: 99%

“…Face alignment refers to detecting facial landmarks such as eye centers, nose tip, and mouth corners. It is the preprocessor stage of many face analysis tasks like face animation [1], face beautification [2], and face recognition [3]. A robust and accurate face alignment is still challenging in unconstrained scenarios, owing to severe occlusions and large appearance variations.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Deep multi-center learning for face alignment

et al. 2020

View full text Add to dashboard Cite

Facial landmarks are highly correlated with each other since a certain landmark can be estimated by its neighboring landmarks. Most of the existing deep learning methods only use one fully-connected layer called shape prediction layer to estimate the locations of facial landmarks. In this paper, we propose a novel deep learning framework named Multi-Center Learning with multiple shape prediction layers for face alignment. In particular, each shape prediction layer emphasizes on the detection of a certain cluster of semantically relevant landmarks respectively. Challenging landmarks are focused firstly, and each cluster of landmarks is further optimized respectively. Moreover, to reduce the model complexity, we propose a model assembling method to integrate multiple shape prediction layers into one shape prediction layer. Extensive experiments demonstrate that our method is effective for handling complex occlusions and appearance variations with real-time performance. The code for our method is available at https://github.com/ZhiwenShao/MCNet-Extension.Index Terms-Multi-Center Learning, Model Assembling, Face Alignment * Corresponding author.(a) Chin is occluded.(b) Right contour is invisible.

show abstract

Section: B Comparison With State-of-the-art Methodsmentioning

confidence: 99%