DualPathGAN: Facial reenacted emotion synthesis

Abstract

Facial reenactment has developed rapidly in recent years, but few methods have been built upon reenacted face in videos. Facial-reenacted emotion synthesis can make the process of facial reenactment more practical. A facial-reenacted emotion synthesis method is proposed that includes a dual-path generative adversarial network (GAN) for emotion synthesis and a residual-mask network to impose structural restrictions to preserve the mouth shape of the source person. To train the dual-path GAN more effectively, a learning strategy based on separated discriminators is proposed. The method is trained and tested on a very challenging imbalanced dataset to evaluate the ability to deal with complex practical scenarios. Compared with general emotion synthesis methods, the proposed method can generate more realistic facial emotion synthesised images or videos with higher quality while retaining the expression contents of the original videos. The DualPathGAN achieves a Fréchet inception distance (FID) score of 9.20, which is lower than the FID score of 11.37 achieved with state-of-the-art methods.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Residual Self-Calibration and Self-Attention Aggregation Network for Crop Disease Recognition

“…In recent years, deep learning techniques, especially convolutional neural networks (CNN) [13][14][15], are rapidly becoming the preferred method to overcome the above-mentioned challenges [16][17][18][19][20]. Due to the scale invariance of the convolutional neural network, the image problem it solves is not limited by the scale and shows outstanding ability in recognition and classification.…”

Bounding‐box deep calibration for high performance face detection

“…Face detection is a fundamental and critical step towards various downstream face applications, such as face alignment [1][2][3], face recognition [4][5][6], and facial attributes synthesis [7][8][9][10] etc. Compared with traditional methods, deep learning methods especially convolutional neural networks (CNNs) have achieved remarkable successes in a variety of computer vision tasks, ranging from image classification [11][12][13][14] to object detection [15][16][17][18][19][20], which also inspire face detection.…”