We develop a gender classification method using convolutional neural networks. We train Alexnet Architecture using the luminance (Y) component of the facial image (YCbCr) for the SoF, groups, and face recognition technology datasets. The Y component is reduced to a size of 32 × 32 via discrete wavelet transform (DWT). The use of the Y plane and a low-resolution subband image of the DWT significantly reduce the amount of processed data. We are able to achieve better results than other machine learning, rule-based approaches and the traditional convolutional neural net structure that are trained with three-dimensional RGB images. We are able to maintain comparably high recognition accuracy, even with the reduction of the number of network layers. We have also compared our structure with the state-of-the-art methods and provided the recognition rates.
We propose an adaptively synchronous scalable spread spectrum (A4S) data-hiding strategy to integrate disparate data, needed for a typical 3-D visualization, into a single JPEG2000 format file. JPEG2000 encoding provides a standard format on one hand and the needed multiresolution for scalability on the other. The method has the potential of being imperceptible and robust at the same time. While the spread spectrum (SS) methods are known for the high robustness they offer, our data-hiding strategy is removable at the same time, which ensures highest possible visualization quality. The SS embedding of the discrete wavelet transform (DWT)-domain depth map is carried out in transform domain YCrCb components from the JPEG2000 coding stream just after the DWT stage. To maintain synchronization, the embedding is carried out while taking into account the correspondence of subbands. Since security is not the immediate concern, we are at liberty with the strength of embedding. This permits us to increase the robustness and bring the reversibility of our method. To estimate the maximum tolerable error in the depth map according to a given viewpoint, a human visual system (HVS)-based psychovisual analysis is also presented.
In this paper we propose a strategy for seamless tessellation, of varying resolution tiles, based on smoothing
and mosaicking in the DWT domain. The scenario involves a tessellation with three different tile qualities
or levels of detail (LOD), at a given instant, depending on the viewpoint distance, the time of rendering and
hardware resources. The LOD is dependent on the multiresolution characteristic of wavelets from the now widely
accepted JPEG2000 codec. Taking the change in viewpoint focus, analogous to a window sliding approach, we
believe that at worst the window may come up to be composed of three different tile qualities with the resultant
artifacts at tile interfaces. To dilute these artifacts, we treat the tiles at the subband level, in the DWT
domain, by employing operations involving suitable subband-sized composite masks, conceived with smoothing
and mosaicking in perspective. The resultant composite subbands are subjected to a global inverse DWT to get
the final seamless tessellation.
The diversity of clients in today's network environment compels us to think about solutions that more than satisfy their needs according to their resources. For 3D terrain visualization this translates into two main requirements, namely the scalability and synchronous unification of a disparate data that requires at least two files, the texture image and its corresponding digital elevation model (DEM). In this work the scalability is achieved through the multiresolution discrete wavelet transform (DWT) of the JPEG2000 codec. For the unification of data, a simple DWT-domain spread spectrum (SS) strategy is employed in order to synchronously hide the DEM in the
corresponding texture while conserving the JPEG2000 standard file format. Highest possible quality texture is renderable due to the reversible nature of the SS data hiding. As far as DEM quality is concerned, it is ensured through the adaptation of synchronization in embedding that would exclude some highest frequency subbands.
To estimate the maximum tolerable error in the DEM according to a given viewpoint, a human visual system (HVS) based psycho-visual analysis is being presented. This analysis is helpful in determining the degree of adaptation in synchronization.
The use of aerial photographs, satellite images, scanned maps and digital elevation models necessitates the
setting up of strategies for the storage and visualization of these data in an interactive way. In order to obtain
a three dimensional visualization it is necessary to map the images, called textures, onto the terrain geometry
computed with Digital Elevation Model (DEM). Practically, all of these informations are stored in three different
files: DEM, texture and geo-localization of the data. In this paper we propose to save all this information in a
single file for the purpose of synchronization. For this, we have developed a wavelet-based embedding method
for hiding the data in a color image. The texture images containing hidden DEM data can then be sent from
the server to a client in order to effect 3D visualization of terrains. The embedding method is integrable with
the JPEG2000 coder to accommodate compression and multi-resolution visualization.
Conference Committee Involvement (3)
First International Workshop on Image Processing Theory, Tools and Applications, IPTA'08
24 November 2008 |
16th European Signal Processing Conference, EUSIPCO'08
25 August 2008 |
12èmes journées d'étude et d'échange COmpression et REprésentation des Signaux Audiovisuels (CORESA 2007)