Ning Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Ning Wang has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
Zhi-Hui Wang
School of Software, Dalian University of Technology, Dalian, China
2 shared publicationsChin-Chen Chang
Department of Information Engineering and Computer Science, Feng Chia University, Taichung, Taiwan
1 shared publicationHairui Yang
School of Software, Dalian University of Technology, Dalian, China
1 shared publicationHaojie Li
School of Software, Dalian University of Technology, Dalian, China
1 shared publicationHong Wang
School of Software, Dalian University of Technology, Dalian, China
1 shared publicationJian Shi
School of Software, Qufu Normal University, Qufu, China
1 shared publication
Latest Publications
- Open Access
- 25 April 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 56683-56693
Multi-Instance Sketch to Image Synthesis With Progressive Generative Adversarial Networks
Real-world images usually contain multiple objects, as a result, generating an image from a multi-instance sketch is an attractive research topic. However, existing generative networks usually produce a similar texture on different instances for those methods focus on learning the distribution of the whole image. To address this problem, we propose a progressive instance texture reserved generative approach to generate more convincible images by decoupling the generation of the instances and the whole image. Specifically, we create an instance generator to synthesize the primitive color distribution and the detailed texture for each instance. Then, an image generator is designed to combine all of these instances to synthesize an image retaining texture and color. Besides, to generate more significant details, such as eyes, ears, and so on, we propose a novel technique called discriminative sketch augmentation, which can provide structural constraint by obtaining the sketch of the discriminative region. The extensive experiments demonstrate that our model not only generates convincing images but also achieves higher inception score and lower Fréchet Inception Distance on the MS-COCO dataset.
- 25 April 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 7, 56683-56693
Open AccessMulti-Instance Sketch to Image Synthesis With Progressive Generative Adversarial Networks
Real-world images usually contain multiple objects, as a result, generating an image from a multi-instance sketch is an attractive research topic. However, existing generative networks usually produce a similar texture on different instances for those methods focus on learning the distribution of the whole image. To address this problem, we propose a progressive instance texture reserved generative approach to generate more convincible images by decoupling the generation of the instances and the whole image. Specifically, we create an instance generator to synthesize the primitive color distribution and the detailed texture for each instance. Then, an image generator is designed to combine all of these instances to synthesize an image retaining texture and color. Besides, to generate more significant details, such as eyes, ears, and so on, we propose a novel technique called discriminative sketch augmentation, which can provide structural constraint by obtaining the sketch of the discriminative region. The extensive experiments demonstrate that our model not only generates convincing images but also achieves higher inception score and lower Fréchet Inception Distance on the MS-COCO dataset.
- Open Access
- 2 April 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 49348-49357
New Secret Sharing Scheme Based on Faster R-CNNs Image Retrieval
Secret image sharing has been attracting considerable research attention in recent years as images become ubiquitous on the Internet and mobile applications. In traditional (t, n) threshold secret image sharing schemes, the sender embeds the secret image into several shadow images and sends them to all participants. However, the shadow images are vulnerable during the transmission process on the Internet. In order to ensure the security of the shadow images which contain secret information, a new secret sharing scheme is proposed based on Faster region convolutional neural networks (Faster R-CNNs). This scheme uses a query image which does not have any secret information but looks similar to the original shadow image. Each participant retrieves the corresponding shadow image to search the query image in a big image database. As a result shadow images are searched by using Faster R-CNNs and stored in a database which is protected from being attacked, rather than transmitted over the network directly in the proposed scheme. The experimental results demonstrate that the proposed scheme can automatically retrieve the accurate shadow images by Faster R-CNNs and recover the secret image correctly.
- 2 April 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 6, 49348-49357
Open AccessNew Secret Sharing Scheme Based on Faster R-CNNs Image Retrieval
Secret image sharing has been attracting considerable research attention in recent years as images become ubiquitous on the Internet and mobile applications. In traditional (t, n) threshold secret image sharing schemes, the sender embeds the secret image into several shadow images and sends them to all participants. However, the shadow images are vulnerable during the transmission process on the Internet. In order to ensure the security of the shadow images which contain secret information, a new secret sharing scheme is proposed based on Faster region convolutional neural networks (Faster R-CNNs). This scheme uses a query image which does not have any secret information but looks similar to the original shadow image. Each participant retrieves the corresponding shadow image to search the query image in a big image database. As a result shadow images are searched by using Faster R-CNNs and stored in a database which is protected from being attacked, rather than transmitted over the network directly in the proposed scheme. The experimental results demonstrate that the proposed scheme can automatically retrieve the accurate shadow images by Faster R-CNNs and recover the secret image correctly.