Adaptive Feature Interpolation for Low-Shot Image Generation

Dai, Mengyu; Hang, Haibin; Guo, X.

doi:10.1007/978-3-031-19784-0_15

Cited by 4 publications

(5 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These issues result in reduced fidelity and unstable training processes. Recent fewshot improvement methods can be divided into the following three categories, i.e., data augmentation (Dai, Hang, and Guo 2022;Jeong and Shin 2021;Karras et al 2020;Zhao et al 2020), regularization (Kong et al 2022;Yang et al 2021;Zhang et al 2019a;Zhao et al 2021), and transfer learning (Mo, Cho, and Shin 2020;Liu et al 2019;Ojha et al 2021). Nevertheless, among prior optimization techniques for few-shot generation, there is still a lack of identity-controllable palmprint generation methods.…”

Section: Few-shot Generationmentioning

confidence: 99%

PCE-Palm: Palm Crease Energy Based Two-Stage Realistic Pseudo-Palmprint Generation

Jin,

Shen,

Zhang

et al. 2024

AAAI

View full text Add to dashboard Cite

The lack of large-scale data seriously hinders the development of palmprint recognition. Recent approaches address this issue by generating large-scale realistic pseudo palmprints from Bézier curves. However, the significant difference between Bézier curves and real palmprints limits their effectiveness. In this paper, we divide the Bézier-Real difference into creases and texture differences, thus reducing the generation difficulty. We introduce a new palm crease energy (PCE) domain as a bridge from Bézier curves to real palmprints and propose a two-stage generation model. The first stage generates PCE images (realistic creases) from Bézier curves, and the second stage outputs realistic palmprints (realistic texture) with PCE images as input. In addition, we also design a lightweight plug-and-play line feature enhancement block to facilitate domain transfer and improve recognition performance. Extensive experimental results demonstrate that the proposed method surpasses state-of-the-art methods. Under extremely few data settings like 40 IDs (only 2.5% of the total training set), our model achieves a 29% improvement over RPG-Palm and outperforms ArcFace with 100% training set by more than 6% in terms of TAR@FAR=1e-6.

show abstract

Section: Few-shot Generationmentioning

confidence: 99%

PCE-Palm: Palm Crease Energy Based Two-Stage Realistic Pseudo-Palmprint Generation

Jin,

Shen,

Zhang

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

“…For the conditional image generation task, it is necessary to introduce additional conditional information to the GAN model to assist the generation, and such information are usually class labels. In this task, the gradient normalization is calculated as shown in Formula (7).…”

Section: Discriminator Architecturementioning

confidence: 99%

“…In recent years, there has been an ongoing intense competition between diffusion models [1-3] and generative adversarial networks (GANs) [4] in various domains, including image generation [5][6][7][8], image super-resolution [9][10][11][12], style transfer [13][14][15][16], image transformation [17][18][19], image-to-image translation [20][21][22], and adversarial attack [23,24]. This competition has greatly pushed and is continuously pushing the development of generative models.…”

Section: Introductionmentioning

confidence: 99%

“…However, these methods greatly increase the workload of the hyperparameter optimization task and the difficulty of training, which may eventually result in unstable training [25,34]. Inspired by recent studies [5,7,14,22,27,42], we believe that by introducing a suitable normalization strategy, it is possible to effectively balance the training stability and image generation quality of GANs. Therefore, in this paper, we try to realize such a GAN model by enhancing the training stability of an existing powerful GAN model (which is with strong image generation ability) through the introduction of a suitable normalization method, while do not compromise the quality of its generated images.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

SUGAN: A Stable U-Net Based Generative Adversarial Network

Cheng,

Wang,

Zhang

et al. 2023

Sensors

View full text Add to dashboard Cite

As one of the representative models in the field of image generation, generative adversarial networks (GANs) face a significant challenge: how to make the best trade-off between the quality of generated images and training stability. The U-Net based GAN (U-Net GAN), a recently developed approach, can generate high-quality synthetic images by using a U-Net architecture for the discriminator. However, this model may suffer from severe mode collapse. In this study, a stable U-Net GAN (SUGAN) is proposed to mainly solve this problem. First, a gradient normalization module is introduced to the discriminator of U-Net GAN. This module effectively reduces gradient magnitudes, thereby greatly alleviating the problems of gradient instability and overfitting. As a result, the training stability of the GAN model is improved. Additionally, in order to solve the problem of blurred edges of the generated images, a modified residual network is used in the generator. This modification enhances its ability to capture image details, leading to higher-definition generated images. Extensive experiments conducted on several datasets show that the proposed SUGAN significantly improves over the Inception Score (IS) and Fréchet Inception Distance (FID) metrics compared with several state-of-the-art and classic GANs. The training process of our SUGAN is stable, and the quality and diversity of the generated samples are higher.

show abstract

“…Compared to other generative models, DCGAN not only enhances the fidelity of generated images but also ensures stable training in deeper networks. However, stable training of the DCGAN requires a substantial dataset, otherwise, issues such as low image fidelity and model collapse may occur [15][16][17]. In order to consistently generate BSTIs, we propose the use of BSTI-DCGAN.…”

Section: Introductionmentioning

confidence: 99%

Space-time image velocimetry in blurred scenes based on BSTI-DCGAN data augmentation

Hu,

Jiang,

Zhang

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Due to the limited sample quantity and the complex data collection process of the blurred space-time image (BSTI) dataset, the deep learning-based space-time image velocimetry (STIV) results in larger errors when applied to blurry videos.. To enhance the measurement accuracy, we propose the use of space-time image velocimetry in blurred scenes based on BSTI-DCGAN data augmentation. Firstly, BSTI-DCGAN is developed based on the deep convolutional generative adversarial network (DCGAN). This network utilizes a bilinear interpolation-convolution module for upsampling and integrates coordinated attention (CA) and multi-concatenation attention (MCA) to enhance the resemblance between generated and real images. Next, further expanding the dataset by using artificially synthesized space-time images.Subsequently, all space-time images are transformed into spectrograms to create a training dataset for the classification network. Finally, the primary spectral direction is detected using the classification network. The experimental results indicate that our approach effectively augments the dataset and improves the accuracy of practical measurements. Under the condition of video blur, the relative errors of the average flow velocity and discharge are 3.92% and 2.72%, respectively.

show abstract

Adaptive Feature Interpolation for Low-Shot Image Generation

Cited by 4 publications

References 22 publications

PCE-Palm: Palm Crease Energy Based Two-Stage Realistic Pseudo-Palmprint Generation

PCE-Palm: Palm Crease Energy Based Two-Stage Realistic Pseudo-Palmprint Generation

SUGAN: A Stable U-Net Based Generative Adversarial Network

Space-time image velocimetry in blurred scenes based on BSTI-DCGAN data augmentation

Contact Info

Product

Resources

About