Self-Sparse Generative Adversarial Networks

Abstract

Generative adversarial networks (GANs) are an unsupervised generative model that learns data distribution through adversarial training. However, recent experiments indicated that GANs are difficult to train due to the requirement of optimization in the high dimensional parameter space and the zero gradient problem. In this work, we propose a self-sparse generative adversarial network (Self-Sparse GAN) that reduces the parameter space and alleviates the zero gradient problem. In the Self-Sparse GAN, we design a self-adaptive sparse transform module (SASTM) comprising the sparsity decomposition and feature-map recombination, which can be applied on multi-channel feature maps to obtain sparse feature maps. The key idea of Self-Sparse GAN is to add the SASTM following every deconvolution layer in the generator, which can adaptively reduce the parameter space by utilizing the sparsity in multi-channel feature maps. We theoretically prove that the SASTM can not only reduce the search space of the convolution kernel weight of the generator but also alleviate the zero gradient problem by maintaining meaningful features in the batch normalization layer and driving the weight of deconvolution layers away from being negative. The experimental results show that our method achieves the best Fréchet inception distance (FID) scores for image generation compared with Wasserstein GAN with gradient penalty (WGAN-GP) on MNIST, Fashion-MNIST, CIFAR-10, STL-10, mini-ImageNet, CELEBA-HQ, and LSUN bedrooms datasets, and the relative decrease of FID is 4.76%-21.84%. Meanwhile, an architectural sketch dataset (Sketch) is also used to validate the superiority of the proposed method.

Feature pyramid network with self-guided attention refinement module for crack segmentation

“…In addition, the proposed refinement module is compared against the Self-Adaptive Sparse Transform Module (SASTM) proposed by Qian et al due to the use of both channel and position attention weights. 67 One difference between the SASTM and the proposed refinement module is that the refinement module uses multi-layer feature maps to refine individual layers of the decoder as compared to inserting the SASTM layer after every deconvolution layer. The SASTM layer is incorporated into networks such as PANet, U-Net and DeepLab v3+ to compare the performance of using SASTM and the proposed refinement module.…”

Application of Voiceprint Recognition Technology Based on Channel Confrontation Training in the Field of Information Security

“…In Equation ( 1), FA is the false acceptance rate, FR is the false rejection rate, fa C denotes the false rejection cost, fr C denotes the false reception cost, Im p P denotes the Figure 1 is a massive voice data resource database, which is mainly composed of telephone voice, Internet voice, and other voice, the three sub-speaker voice databases. In storage facing massive speaker representation, distributed storage is the interface for storing external file upload, file download, file modification, and file deletion [16]. Distributed storage is to create the voiceprint sublibrary according to the gender, age, and other information of the speaker, which is the most concise and effective way to build the voiceprint word library.…”