Improving Face Sketch Recognition via Adversarial Sketch-Photo Transformation

Yu, Shikang; Han, Hu; Shan, Shiguang; Dantcheva, Antitza; Chen, Xilin

doi:10.1109/fg.2019.8756563

Cited by 18 publications

(13 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The intra-modal methods [8], [22] firstly reduce the modality gap by synthesizing a sketch into a photo and then use traditional face recognition methods to match the synthetic photo with the original photo. Under the assumption that the synthetic process can be treated as a linear mapping, Tang et al [26] used principal component analysis (PCA) to reconstruction coefficients from training photos firstly, and then they synthesized the sketch by the learned reconstruction coefficients.…”

Section: Related Workmentioning

confidence: 99%

“…With the development of deep learning, Generative Adversarial Networks (GAN) [28] are recently used to generate a synthetic photo, they belong to intra-modality methods and have achieved good performance. Yu et al [22] proposed a conditional CycleGAN to obtain the synthetic photo. Specifically, they designed a feature-level loss to guide the network to generate highquality synthetic photos.…”

Section: Related Workmentioning

confidence: 99%

“…metric learning methods have been widely utilized for face recognition and other recognition problems [16]- [19]. The CNNs exhibit expressive power, so there has been a growing interest in using CNNs to learn discriminative features for sketch face recognition [10], [20]- [22]. However, the success of deep learning relies on the largescale paired images to counteract effects such as over-fitting [23].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Domain Alignment Embedding Network for Sketch Face Recognition

et al. 2021

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Domain Alignment Embedding Network for Sketch Face Recognition

et al. 2021

View full text Add to dashboard Cite

“…Baselines for Sketch Face Recognition We compare against various state-of-the-art baselines for sketch face recognition, including CAL-HFR [25], DVR [50], DLFace [39], LightCNN+DVG [14], IACycleGAN [12], ASPT [51], and RCN [11].…”

Section: Sketch Face Recognitionmentioning

confidence: 99%

Learning from the Past: Meta-Continual Learning with Knowledge Embedding for Jointly Sketch, Cartoon, and Caricature Face Recognition

Zheng

Lan

Wang

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

View full text Add to dashboard Cite

This paper deals with a challenging task of learning from different modalities by tackling the difficulty problem of jointly face recognition between abstract-like sketches, cartoons, caricatures and real-life photographs. Due to the significant variations in the abstract faces, building vision models for recognizing data from these modalities is an extremely challenging. We propose a novel framework termed as Meta-Continual Learning with Knowledge Embedding to address the task of jointly sketch, cartoon, and caricature face recognition. In particular, we firstly present a deep relational network to capture and memorize the relation among different samples. Secondly, we present the construction of our knowledge graph that relates image with the label as the guidance of our meta-learner. We then design a knowledge embedding mechanism to incorporate the knowledge representation into our network. Thirdly, to mitigate catastrophic forgetting, we use a meta-continual model that updates our ensemble model and improves its prediction accuracy. With this meta-continual model, our network can learn from its past. The final classification is derived from our network by learning to compare the features of samples.

show abstract

“…Li et al [15] as well as Jo and Park [16] generated facial images with a partial reconstruction from sketches. In terms of application, our work is related to Lu et al [17] and Yu et al [18]. The former uses Contextual GAN, where input photos and images are trained in semi-supervised fashion.…”

Section: Related Workmentioning

confidence: 99%

Generator From Edges: Reconstruction of Facial Images

Takano¹,

Alaghband²

2020

Preprint

View full text Add to dashboard Cite

Applications that involve supervised training require paired images. Researchers of single image superresolution (SISR) create such images by artificially generating blurry input images from the corresponding ground truth. Similarly we can create paired images with the canny edge. We propose Generator From Edges (GFE) [Figure 2]. Our aim is to determine the best architecture for GFE, along with reviews of perceptual loss [1,2]. To this end, we conducted three experiments. First, we explored the effects of the adversarial loss often used in SISR. In particular, we uncovered that it is not an essential component to form a perceptual loss. Eliminating adversarial loss will lead to a more effective architecture from the perspective of hardware resource. It also means that considerations for the problems pertaining to generative adversarial network (GAN) [3], such as mode collapse, are not necessary. Second, we reexamined VGG loss and found that the mid-layers yield the best results. By extracting the full potential of VGG loss, the overall performance of perceptual loss improves significantly. Third, based on the findings of the first two experiments, we reevaluated the dense network to construct GFE. Using GFE as an intermediate process, reconstructing a facial image from a pencil sketch can become an easy task.

show abstract

Improving Face Sketch Recognition via Adversarial Sketch-Photo Transformation

Cited by 18 publications

References 33 publications

Domain Alignment Embedding Network for Sketch Face Recognition

Domain Alignment Embedding Network for Sketch Face Recognition

Learning from the Past: Meta-Continual Learning with Knowledge Embedding for Jointly Sketch, Cartoon, and Caricature Face Recognition

Generator From Edges: Reconstruction of Facial Images

Contact Info

Product

Resources

About