【人脸对齐-Landmarks】人脸关键点检测方法及评测汇总
生活随笔
收集整理的這篇文章主要介紹了
【人脸对齐-Landmarks】人脸关键点检测方法及评测汇总
小編覺得挺不錯的,現在分享給大家,幫大家做個參考.
好東西往往都在最后面,請耐心向下滑
傳統方法:
1. SDM:Supervised descent method and its applications to face alignment. CVPR2013
速度:
CPU(i7) - 小于12ms精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (5.60), challenge: (15.40), Full: (7.52).評價:
2. ERT:One millisecond face alignment with an ensemble of regression trees. CVPR2014
速度:
CPU(i7) - 小于6ms精度:
評測數據集1---300w (Inter-pupil Normalisation): common: Full: (6.40).評價:
3. LBF:Face alignment at 3000 fps via regressing local binary features, 2014
速度:
CPU(i5) - PC上至少300fps,最快3000fps精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (4.95), challenge: (11.98), Full: (6.32). 評測數據集2---IBUG: 11.98評測數據集3---LFPW: 3.35評測數據集4---HELEN: 5.41評價:
4. CFSS:Face Alignment by Coarse-to-Fine Shape Searching, 2015
速度:
CPU(i5) - 25fps精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (4.73), challenge: (9.98), Full: (5.76).評價:
深度方法:
1. cascade CNN:Deep Convolutional Network Cascade for Facial Point Detection,2013
速度:
GTX1070 - 8msCPU - 120ms精度:
待補充評價:
2. Face++:Extensive facial landmark localization with coarse-to-fine convolutional network cascade,2013
速度: 未知
精度:
評價:
沒有與其他方法對比,此方法效果稍微差點。3. TCDCN:Learning and Transferring Multi-task Deep Representation for Face Alignment,2014
速度:
GTX760 - 1.5msCPU(i5) - 17ms精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (4.80), challenge: (8.60), Full: (5.54). 評測數據集2---IBUG: 9.15評價:
4. MCSR:M3 CSR: Multi-view, multi-scale and multi-component cascade,2016
速度:
CPU(i7) - 50ms精度:
評測數據集2---IBUG: 5.65評價:
5. Approaching human level facial landmark localization by deep learning,2016
速度:
CPU(i7) - 500ms精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (3.43), challenge: (5.72), Full: (3.88). 評測數據集2---IBUG: 5.65評價:
6. MDM:Mnemonic Descent Method:A recurrent process applied for end-to-end face alignment, 2016
速度: 未知
精度:
評價:
7. Unconstrained Face Alignment via Cascaded Compositional Learning,2016
速度:
CPU - 350FPS精度:
評測數據集3---AFLW : 待補充評價:
8. DRDA:Occlusion-free Face Alignment: Deep Regression Networks Coupled with De-corrupt AutoEncoders,2016
速度:
未知精度:
評測數據集2---IBUG: 待補充評價:
9. Stacked Hourglass Network for Robust Facial Landmark Localisation, CVPR 2017
速度:
未知精度:
評測數據集1---300w (Inter-pupil Normalisation ):paper: common: 4.12,challenge: 7.00, Full: -. 復現: common: 3.84,challenge: 7.37, Full:4.54.評價:
開源:
10. Style Aggregated Network for Facial Landmark Detection, CVPR 2018
速度:
未知精度:
待補充評價:
待補充
開源:
11. Wing Loss for RobustFacial Landmark Localisation with Convolutional Neural Networks. 2018
速度:
GTX1080 TITAN - 60msCPU - 100/15/5 FPS精度:
評測數據集1---300w (Inter-pupil Normalisation ):common: 3.27,challenge: 7.18, Full: 4.04.評價:
12. LAB:Look at Boundary: A Boundary-Aware Face Alignment Algorithm. CVPR2018.
速度:
GTX1080 TITAN - 60msCPU - 未知精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (4.20, 3.42, 2.57), challenge: (7.41, 6.98, 4.72), Full: (4.92, 4.12, 2.99). 評測數據集1.2---300w (Inter-ocular Normalisation):common: (2.98, 2.43, 1.85), challenge: (5.19, 4.85, 3.28), Full: (3.49, 2.93, 2.13).評價:
13. DCFE:A Deeply-initialized Coarse-to-fine Ensemble of Regression Trees for Face Alignment. 2018
速度:
環境:NVidia GeForce GTX 1080 (8GB) GPU Intel Xeon E5-1650 at 3.50GHz (6 cores/12 threads, 32 GB of RAM)C++, Tensorflow , OpenCV libraries.訓練時間:48h預測速度:GTX 1080 (8GB) - 32 FPS,其中 CNN 25ms,ERT-6.25msCPU - 未知精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (3.83), challenge: (7.54), Full: (4.55). 評測數據集1.3---300w (Inter-corners Normalisation):common: (2.76), challenge: (5.22), Full: (3.24).評價:
比LAB低,但速度比LAB快
14. PFLD:A Practical Facial Landmark Detector. 2019
速度:
ARM 845 140fps精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (3.17), challenge: (6.33), Full: (3.76). 評測數據集1.3---300w (Inter-ocular Normalisation):common: (2.96), challenge: (4.98), Full: (3.37).評價:
開源:
15. ODN:Robust Facial Landmark Detection via Occlusion-adaptive Deep Networks. CVPR 2019
速度:
待補精度:
評測數據集1.3---300w (Inter-ocular Normalisation):common: (3.56 ), challenge: (6.67), Full: (4.17).評價:
wing loss + LAB
開源:
16. AWing:Adaptive Wing Loss for Robust Face Alignment via Heatmap Regression
速度:
待補精度:
評測數據集1---300w (Inter-pupil Normalisation): common: (3.77 ), challenge: (6.32), Full: (4.31). 評測數據集1.3---300w (Inter-ocular Normalisation):common: (2.72), challenge: (4.52), Full: (3.07).評價:
待補
開源:
17. LUVLi Face Alignment: Estimating Landmarks’ Location, Uncertainty, and Visibility Likel
速度:
精度:
評測數據集1.3---300w (Inter-ocular Normalisation):common: (2.76), challenge: (5.16), Full: (3.23).評價:
開源:
18. ATF: Towards Robust Face Alignment via Leveraging Similarity and Diversity across Different Datasets
速度:
精度:
評測數據集1.3---300w (Inter-ocular Normalisation):common: (2.76), challenge: (5.16), Full: (3.23).評價:
開源:
指標匯總:
1. 數據集 300w
2. 數據集 WFLW
To Do List:
參考博文:
[1] https://blog.csdn.net/xzzppp/article/details/74939823
[2] https://blog.csdn.net/Taily_Duan/article/details/54376467
[3] https://blog.csdn.net/Mynameisyournamewuyu/article/details/85490059
[4] https://blog.csdn.net/wwwhp/article/details/88361422
總結
以上是生活随笔為你收集整理的【人脸对齐-Landmarks】人脸关键点检测方法及评测汇总的全部內容,希望文章能夠幫你解決所遇到的問題。
- 上一篇: Github: 从github上拉取别人
- 下一篇: IDC 研究:2022 中国折叠屏手机市