Entity

Time filter

Source Type


Wu Y.,Nanjing University of Aeronautics and Astronautics | Wu Y.,State Key Laboratory of Transient Optics and Photonics | Yin J.,Nanjing University of Aeronautics and Astronautics | Zhu L.,Nanjing University of Aeronautics and Astronautics | Ye Z.,Nanjing University of Aeronautics and Astronautics
Zhongguo Jiguang/Chinese Journal of Lasers | Year: 2014

Reconstructed image of digital holography has problems such as serious interference of speckle noise, low contrast and so on. Thus a method based on speckle reducing anisotropic diffusion (SRAD) model and nonsubsampled contourlet transform (NSCT) is proposed for improving reconstructed image quality of digital hologram. SRAD model is adopted to eliminate speckle noise of reconstructed image. After NSCT decomposition, a low-frequency sub-band and several high-frequency sub-bands are produced. Low-frequency sub-band coefficients are adjusted based on a nonlinear gain function and an image segmentation method. In high-frequency sub-bands, edges are enhanced using a NSCT modulus maximum edge detection method. A large number of experimental results show that, compared with nonlinear diffusion denoising methods and NSCT enhancement methods proposed recently, the proposed method can more effectively eliminate speckle noise and improve the contrast of reconstructed image. Furthermore, the edges are smooth and clear. As a result, the accuracy of recognition and measurement in digital holography can be improved. Source


Ji R.,Xiamen University | Gao Y.,Tsinghua University | Liu W.,IBM | Xie X.,Microsoft | And 2 more authors.
ACM Transactions on Intelligent Systems and Technology | Year: 2015

Coming with the popularity of multimedia sharing platforms such as Facebook and Flickr, recent years have witnessed an explosive growth of geographical tags on social multimedia content. This trend enables a wide variety of emerging applications, for example, mobile location search, landmark recognition, scene reconstruction, and touristic recommendation, which range from purely research prototype to commercial systems. In this article, we give a comprehensive survey on these applications, covering recent advances in recognition and mining of geographical-aware social multimedia. We review related work in the past decade regarding to location recognition, scene summarization, tourism suggestion, 3D building modeling, mobile visual search and city navigation. At the end, we further discuss potential challenges, future topics, as well as open issues related to geo-social multimedia computing, recognition, mining, and analytics. © 2015 ACM. Source


Wu Y.,Nanjing University of Aeronautics and Astronautics | Wu Y.,State Key Laboratory of Transient Optics and Photonics | Ye Z.,Nanjing University of Aeronautics and Astronautics | Wan H.,Nanjing University of Aeronautics and Astronautics
Zhongguo Jiguang/Chinese Journal of Lasers | Year: 2013

In order to suppress the speckle noise of reconstructed image in digital holography effectively and to further improve the quality of reconstructed image, a speckle noise suppression method of reconstructed image in digital holography based on dual-tree complex wavelet transform (DT-CWT) and anisotropic diffusion is proposed. The reconstructed image is decomposed through DT-CWT. Then, the low-frequency components and the high-frequency components in six directions are processed through P_Laplace diffusion and Laplacian pyramid-based nonlinear diffusion (LPND), respectively. The new reconstructed image is synthesized through inverse dual-tree complex wavelet transform (IDT-CWT). The experimental results are given, and a subjective visual comparison is made with the method of wavelet threshold shrinkage and total variation (TV) diffusion, the method of LPND, and the method combining contourlet transform with TV and adaptive contrast diffusion. While the results are evaluated quantitatively according to peak signal to noise ratio (PSNR), correlation coefficient (COR) and running time. Experimental results show that the proposed method has a better performance in speckle noise suppression and preserves the detail and textural features of original reconstructed image more efficiently. Source


Wu Y.,Nanjing University of Aeronautics and Astronautics | Wu Y.,State Key Laboratory of Transient Optics and Photonics | Yin J.,Nanjing University of Aeronautics and Astronautics
Zhongguo Jiguang/Chinese Journal of Lasers | Year: 2013

Aiming at the problem of low contrast, blurred edges and textures, and speckle noise of reconstructed image in digital holography, an adaptive enhancement method for reconstructed image of digital holography in contourlet domain based on uniform searching particle swarm optimization is proposed. Median filtering algorithm is used to suppress speckle noise of the reconstructed image. After the contourlet decomposition, edge enhancement is performed for the band-pass directional subbands by a nonlinear gain function. While the coefficients of low-pass subbands are adjusted by the gain function based on a gray-scale transform and the local mean. The gray-scale transform aims to expand the dark areas of digital holographic image. The undetermined parameters are found by uniform searching particle swarm optimization. The fitness function takes into account the contrast, definition and peak signal-to-noise ratio of image. A large number of experimental results show that, compared with three existing enhancement methods, the proposed method can more effectively improve the contrast and definition of reconstructed image in digital holography, highlight edges and textures, and suppress speckle noise. As a result, the measurement accuracy of digital holography can be improved. Source


Zhang Y.F.,Nanjing University of Aeronautics and Astronautics | Wu Y.Q.,State Key Laboratory of Transient Optics and Photonics
Advanced Materials Research | Year: 2013

In the reconstruction process of digital holograph, zero-order image has a bad impact on the quality of real image. In this paper, a new way to eliminate zero-order image in digital holography is proposed. Firstly, digital holography image is decomposed by contourlet transform, then remove the low frequency. The new digital holography image is obtained by inverse contourlet transform. Experiments show that, compared with spatial filtering, frequency domain filtering, laplacian filtering and eliminate zero-order image method based on wavelet, the new method proposed in this paper can eliminate zero-order image better, what's more, the real image is also strengthened in some way. © (2013) Trans Tech Publications, Switzerland. Source

Discover hidden collaborations