隨著人臉識別被廣泛運用在各種安全關鍵應用中，Face Anti-Spoofing(FAS)
的相關研究也越來越受關注。FAS 在保護人臉識別系統免於受到 Presentation
Attacks(PAs) 的方面，有著至關重要的作用。然而，很少有 FAS 方法參考人
臉圖片質量，並且很少考慮在不受約束的場景下捕獲足夠高質量圖像時遇到
的困難。以前的方法沒有考慮對 inputed frame 計算質量分數，也沒有根據圖
像質量對不同 frame 應用不同的權重來做出反欺騙的最終決策，因此未能利
用高質量的圖像做出更準確的預測結果。為了解決這個問題，我們開發了基
於人臉圖像質量評估的人臉反欺騙系統。預期的貢獻是，在提取特徵向量
時，質量評估模塊生成輸入圖像的質量分數，透過更依賴質量好的資料，更
好地去區分真人臉圖像和假人臉圖像。我們通過對兩個公共數據集的實驗比
較，來展示所提出的基於 FIQA 的 FAS 系統的表現。

With the widespread use of face recognition in various security-critical applications,
research on face anti-spoofing (FAS) has attracted increasing attention. Face antispoofing plays a vital role in protecting face recognition systems from PAs. However, there are few FAS methods that refer to face image quality and rarely take
into account the difficulties encountered in capturing high-quality images under unconstrained scenarios. Previous methods do not consider computing quality scores
to the input frames or apply different weighting to different frames based on the
image quality to make the final decision for anti-spoofing, thus they failed to exploit the high-quality images to make more accurate prediction result. To solve this
problem, we developed a Face Image Quality Assessment (FIQA) based face antispoofing system. The expected contribution is that while extracting feature vectors,
the quality assessment module generated the quality scores for the input images to
make better decisions on distinguishing real face images from fake face images. We
demonstrate the proposed FIQA-based face anti-spoofing system through experimental comparisons on two public datasets.

[1] Best-Rowden, L., and Jain, A. K. Automatic face image quality prediction,
2017.
[2] Biometrics., I. J. S. Information technology–biometric presentation attack
detection–part 3: testing and reporting.
[3] Boulkenafet, Z., Komulainen, J., and Hadid, A. Face anti-spoofing based on
color texture analysis. In IEEE International Conference on Image Processing
(ICIP), 2015 (2015), IEEE, pp. 2636–2640.
[4] Boulkenafet, Z., Komulainen, J., and Hadid, A. Face spoofing detection using colour texture analysis. IEEE Transactions on Information Forensics and
Security 11, 8 (Aug 2016), 1818–1830.
[5] Boulkenafet, Z., Komulainen, J., Li, L., Feng, X., and Hadid, A. Oulu-npu: A
mobile face presentation attack database with real-world variations. In 2017
12th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2017) (2017), pp. 612–618.
[6] Chingovska, I., Anjos, A., and Marcel, S. On the effectiveness of local binary
patterns in face anti-spoofing. In 2012 BIOSIG - Proceedings of the International Conference of Biometrics Special Interest Group (BIOSIG) (2012),
pp. 1–7.
[7] Feng, H., Hong, Z., Yue, H., Chen, Y., Wang, K., Han, J., Liu, J., and Ding, E.
Learning generalized spoof cues for face anti-spoofing, 2020.
[8] Hernandez-Ortega, J., Galbally, J., Fierrez, J., Haraksim, R., and Beslay, L.
Faceqnet: Quality assessment for face recognition based on deep learning,
2019.
[9] Jia, Y., Zhang, J., Shan, S., and Chen, X. Single-side domain generalization
for face anti-spoofing, 2020.
[10] Khammari, M. Robust face anti-spoofing using cnn with lbp and wld. IET
Image Processing 13, 11 (2019), 1880–1884.
[11] Kim, M., Jain, A. K., and Liu, X. Adaface: Quality adaptive margin for face
recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision
and Pattern Recognition (2022).
[12] King, D. E. Dlib-ml: A machine learning toolkit. The Journal of Machine
Learning Research 10 (2009), 1755–1758.
[13] Li, L., Feng, X., Boulkenafet, Z., Xia, Z., Li, M., and Hadid, A. An original
face anti-spoofing approach using partial convolutional neural network. In
2016 Sixth International Conference on Image Processing Theory, Tools and
Applications (IPTA) (2016), pp. 1–6.
[14] Lin, T.-Y., Goyal, P., Girshick, R., He, K., and Dollár, P. Focal loss for dense
object detection, 2017.
28
[15] Liu, R., and Tan, W. Eqface: A simple explicit quality network for face recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and
Pattern Recognition (CVPR) Workshops (June 2021), pp. 1482–1490.
[16] Liu, Y., Jourabloo, A., and Liu, X. Learning deep models for face antispoofing: Binary or auxiliary supervision, 2018.
[17] Liu, Y., Stehouwer, J., Jourabloo, A., and Liu, X. Deep tree learning for zeroshot face anti-spoofing, 2019.
[18] Liu, Y., Stehouwer, J., and Liu, X. On disentangling spoof trace for generic
face anti-spoofing, 2020.
[19] Määttä, J., Hadid, A., and Pietikäinen, M. Face spoofing detection from single
images using micro-texture analysis. In 2011 International Joint Conference
on Biometrics (IJCB) (2011), pp. 1–7.
[20] Ou, F.-Z., Chen, X., Zhang, R., Huang, Y., Li, S., Li, J., Li, Y., Cao, L.,
and Wang, Y.-G. Sdd-fiqa: Unsupervised face image quality assessment with
similarity distribution distance. In Proceedings of the IEEE/CVF Conference
on Computer Vision and Pattern Recognition (CVPR) (June 2021), pp. 7670–
7679.
[21] Paszke, A., Gross, S., Chintala, S., Chanan, G., Yang, E., DeVito, Z., Lin, Z.,
Desmaison, A., Antiga, L., and Lerer, A. Automatic differentiation in pytorch.
[22] Patel, K., Han, H., and Jain, A. K. Secure face unlock: Spoof detection on
smartphones. IEEE Transactions on Information Forensics and Security 11,
10 (2016), 2268–2283.
[23] Rao, Y., Lu, J., and Zhou, J. Attention-aware deep reinforcement learning for
video face recognition. In 2017 IEEE International Conference on Computer
Vision (ICCV) (2017), pp. 3951–3960.
[24] Rehman, Y. A. U., Po, L.-M., and Komulainen, J. Enhancing deep discriminative feature maps via perturbation for face presentation attack detection. Image
and Vision Computing 94 (2020), 103858.
[25] Shi, Y., and Jain, A. K. Probabilistic face embeddings, 2019.
[26] Terhörst, P., Kolf, J. N., Damer, N., Kirchbuchner, F., and Kuijper, A. Ser-fiq:
Unsupervised estimation of face image quality based on stochastic embedding
robustness, 2020.
[27] van der Maaten, L., and Hinton, G. Viualizing data using t-sne. Journal of
Machine Learning Research 9 (11 2008), 2579–2605.
[28] Wang, Y.-C., Wang, C.-Y., and Lai, S.-H. Disentangled representation with
dual-stage feature learning for face anti-spoofing. In Proceedings of the
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)
(January 2022), pp. 1955–1964.
29
[29] Wen, D., Han, H., and Jain, A. K. Face spoof detection with image distortion analysis. IEEE Transactions on Information Forensics and Security 10, 4
(2015), 746–761.
[30] Zhang, Z., Yan, J., Liu, S., Lei, Z., Yi, D., and Li, S. Z. A face antispoofing
database with diverse attacks. In 2012 5th IAPR International Conference on
Biometrics (ICB) (2012), pp. 26–31.