16 July 2020 in IEEE Access
Image translation tasks based on generative models have become an important research area, such as the general framework for unsupervised image translation-CycleGAN (Cycle-Consistent Generative Adversarial Networks). A typical advantage of CycleGAN is that it can realize the training of two image sets without pairing, but there are still some problems in the preservation of semantic information and the learning of specific features. In this paper, we propose the CycleGAN-AdaIN framework based on the CycleGAN model, which can translate real photos into Chinese ink paintings. In order to retain the content of the image completely, we use one cycle consistency loss to replace two in the structure of the model. To learn the style information of the ink painting, we introduce an AdaIN (Adaptive Instance Normalization) module before the decoding process of the generation network. In addition, to correct the details of the generated image, we add the MS-SSIM (Multi-Scale-Structural Similarity Index) loss in the reconstruction loss to generate a higher quality image. Compared with the existing methods in FID, Kernel MMD, PSNR and SSIM, the experiment results show that our method can accomplish the task of transferring real photos to ink paintings and get better performance than the baseline model.
4 June 2021 in IEEE Transactions on Neural Networks and Learning Systems
A Bayesian nonparametric approach for estimation of a Dirichlet process (DP) mixture of generalized inverted Dirichlet distributions [i.e., an infinite generalized inverted Dirichlet mixture model (InGIDMM)] has been proposed. The generalized inverted Dirichlet distribution has been proven to be efficient in modeling the vectors that contain only positive elements. Under the classical variational inference (VI) framework, the key challenge in the Bayesian estimation of InGIDMM is that the expectation of the joint distribution of data and variables cannot be explicitly calculated. Therefore, numerical methods are usually applied to simulate the optimal posterior distributions. With the recently proposed extended VI (EVI) framework, we introduce lower bound approximations to the original variational objective function in the VI framework such that an analytically tractable solution can be derived. Hence, the problem in numerical simulation has been overcome. By applying the DP mixture technique, an InGIDMM can automatically determine the number of mixture components from the observed data. Moreover, the DP mixture model with an infinite number of mixture components also avoids the problems of underfitting and overfitting. The performance of the proposed approach is demonstrated with both synthesized data and real-life data applications.
1 April 2021 in Neurocomputing
25 October 2021 in International Journal of Intelligent Systems
Bayesian estimation of parameters in the Dirichlet mixture process of the Beta-Liouville distribution (i.e., the infinite Beta-Liouville mixture model) has recently gained considerable attention due to its modeling capability for proportional data. However, applying the conventional variational inference (VI) framework cannot derive an analytically tractable solution since the variational objective function cannot be explicitly calculated. In this paper, we adopt the recently proposed extended VI framework to derive the closed-form solution by further lower bounding the original variational objective function in the VI framework. This method is capable of simultaneously determining the model's complexity and estimating the model's parameters. Moreover, due to the nature of Bayesian nonparametric approaches, it can also avoid the problems of underfitting and overfitting. Extensive experiments were conducted on both synthetic and real data, generated from two real-world challenging applications, namely, object detection and text categorization, and its superior performance and effectiveness of the proposed method have been demonstrated.