다중 측면 얼굴 이미지를 이용한 개선된 얼굴 정면화 기법
Improved Face Frontalization Using Multiple Side-Face Images
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
측면 얼굴 사진을입력으로 정면 얼굴을 생성하는 연구는 얼굴 인식 분야에서 중요한 주제 중 하나이다. 이를 해결하기 위 해, 본 연구에서는 여러 측면 얼굴 이미지를 입력으로 활용하여 성능을 향상시키는 방법을 제안하였다. 여러 각도의 얼굴 이미지를 사용함으로써, 특히 90도와 같은 큰 각도에서도 정확하게 정면 얼굴을재구성할 수 있는능력을 향상시켰다. 다중 이미지를 사용함으로써 단일이미지 방식에서 놓칠 수 있는 세부 정보를 포착할 수 있어, 더 정확하고 신원 보존적인 정면 이미지를 생성할 수 있었다. 실험 결과, 본 연구의 방법이 기존 방식보다 성능이 크게 향상되었으며, 특히 극단적인 각도를 처리할 때 뛰어난 성능을 보였다.
Abstract
The task of generating frontal face images is a crucial challenge in the field of facial recognition. To address this, we propose a method that improves performance by leveraging multiple side-face images as input. By using multiple views of the face, our approach enhances the ability to accurately reconstruct the frontal face, even in cases with large pose variations, such as 90-degree angles. The use of multiple images allows for the capture of details that are otherwise missed in single-image methods, leading to more precise and identity-preserving frontal images. Experimental results show that our method significantly outperforms existing approaches, especially in handling extreme pose angles, demonstrating its effectiveness in improving face frontalization performance.
참고문헌 (11)
- Y. Qian, "Unsupervised face normalization with extreme pose and expression in the wild", Proceedings of the IEEE, 2019, 9851~9858
- R. Huang, "Beyond face rotation : Global and local perception gan for photorealistic and identity preserving frontal view synthesis", Proceedings of the IEEE international conference on computer vision, pp.2439~2448 (2017)
- L. Tran, "Disentangled representation learning gan for pose-invariant face recognition", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.1415~1424 (2017)
- V. Blanz, "A morphable model for the synthesis of 3d faces", Proceedings of the 26th Annual Conference on Computer Graphics and Interactive Techniques, ser. SIGGRAPH ’99, pp.187~194 (1999), ACM Press[DOI]
- Q. Cao, "Vggface2 : A dataset for recognising faces across pose and age", 2018 13th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018), pp.67~74 (2018), IEEE Press
- J. Johnson, "Computer Vision – ECCV 2016", pp.694~711 (2016), Springer International Publishing
- I. Goodfellow, "Generative adversarial nets", Advances in neural information processing systems, 27 (2014)
- M. Arjovsky, "Wasserstein generative adversarial networks", International conference on machine learning, pp.214~223 (2017), PMLR
- I. Gulrajani, "Improved training of wasserstein gans", Advances in neural information processing systems, 30 (2017)
- R. Gross, "Multi-pie", Image and vision computing, 28(5), pp.807~813 (2010)
- K. Zhang, "Joint face detection and alignment using multitask cascaded convolutional networks", IEEE Signal Processing Letters, 23(10), pp.1499~1503 (2016)[DOI]