Ramesh Kumar Lama

Color image interpolation for high resolution display based on adaptive directional lifting based wavelet transform

We propose a new method for the interpolation of full high definition (HD) image based on the adaptive directional wavelet transform (ADWT). The ADWT is the lifting implementation of wavelet transform and is able to transform an image along diagonal orientation as well as conventional horizontal and vertical orientations. In the proposed method the high-frequency sub band images are obtained by ADWT and the input image are interpolated, followed by combining these all images to generate a new interpolated image by using the inverse ADWT. Simulation results demonstrate that our new interpolation algorithm substantially improves the subjective as well as objective quality of the interpolated images over conventional linear interpolation and edge directed interpolations.

Multimedia digital rights management based on selective encryption for flexible business model

This paper proposes an advanced encryption of MP3 and MPEG-4 coder with a quality degradation-based security model. For the MP3 audio, the magnitude and phase information of modified discrete cosine transform (MDCT) coefficients is encrypted. DCT coefficients and motion vectors (MVs) are used for the scrambling of the MPEG-4 video. This encryption scheme has a level of security, secures in perception, keeps format compliance, and obtains better time efficiency by reducing the encrypted volumes of multimedia contents. These properties make it practical to incorporate encryption and decryption process into compression and decompression process, and thus suitable for secure A/V transmission or sharing. Experimental results indicate that the proposed technique provides levels of security for a flexible business model and achieves a simple and coding-efficient architecture with no adverse impact on error resilience.

New interpolation method based on combination of Discrete cosine transform and wavelet transform

In this paper, we propose an image interpolation algorithm that uses a hybrid transform and zero padding scheme. The proposed transform uses the Discrete Wavelet Transform to decompose the image into multi resolution subbands where the Discrete cosine transform (DCT) is used to interpolate the image. Then we use Zeropadding method to interpolate the DCT transformed high frequency wavelet coefficients. The upscaled image is generated using the inverse DWT of interpolated coefficients and the original image. The experimental results demonstrate the efficiency of the proposed hybrid method in terms of peak signal-to-noise ratio PSNR and visual quality.

SVD based improved secret fragment visible mosaic image generation for information hiding

This paper proposes a new method for information hiding based on secret fragment visible mosaic image. This algorithm incorporates the application of color transfer for the mosaic image generation to hide the secret information. Since the RGB color space has strong correlation between three color channels, the modification process of any color is complicated in this color space. The large size of bit stream of recovery information of secret image block sequence creates an additional problem. We present a new way to solve these problems using the color transfer technique in correlated color space. The pixel value is considered as a three dimensional stochastic variable and image as a set of samples. Color transformation is achieved by the geometrical transformations such as translation, scaling and rotation. The correlation between the color spaces is measured and further used for the color transfer process by exploring the covariance. In order to get the robust encryption and lower bit rate requirement, the combination of logistic map and Chebyshev map is used to generate the encryption sequence of the secret image.

Interpolation using wavelet transform and discrete cosine transform for high resolution display

We propose a new interpolation method based on the hybrid technique combining the discrete wavelet transform (DWT) and discrete cosine transform (DCT). The high frequency wavelet coefficients are interpolated using the zero pad method in the DCT domain. The upscaled image is generated using inverse DWT of interpolated coefficients and original image.

Image interpolation for high-resolution display based on the complex dual-tree wavelet transform and hidden Markov model

This paper presents a new method for the interpolation of a full high-definition (HD) image based on the dual-tree complex wavelet transform (DT-CWT) and hidden markov model (HMM). In the proposed method, the DT-CWT is used to decompose the low-resolution image into different subbands. In wavelet domain interpolation, given image is assumed as the low frequency LL subband of the wavelet coefficients of a high-resolution image. The proposed method estimates the higher band coefficients by learning the correlation between the coefficients across the scale. In this paper, the relationship between the wavelet coefficients across the scale is described by HMM, and each wavelet coefficient is modeled by a Gaussian mixture having multiple means and variances. Experimental results show that the proposed algorithm yields images that are sharper compared to several other methods that we have considered in this paper.

Multiple Description Coding Using Directional Discrete Cosine Transform

Delivery of high quality video over a wide area network with large number of users poses great challenges for the video communication system. To ensure video quality, multiple descriptions have recently attracted various attention as a way of encoding and visual information delivery over wireless network. We propose a new efficient multiple description coding (MDC) technique. Quincunx lattice sub-sampling is used for generating multiple descriptions of an image. In this paper, we propose the application of a directional discrete cosine transform (DCT) to a sub-sampled quincunx lattice to create an MDC representation. On the decoder side, the image is decoded from the received side information. If all the descriptions arrive successfully, the image is reconstructed by combining the descriptions. However, if only one side description is received, decoding is executed using an interpolation process. The experimental results show that such the directional DCT can achieve a better coding gain as well as energy packing efficiency than the conventional DCT with re-alignment.

Post processing algorithm for reduction of blocking artifact in HDTV

In this paper, we propose a new post-processing method to reduce the blocking artifacts in digital high definition television (HDTV) decoded images. This algorithm performs the deblocking operation in two modes that are determined by number of edge pixels around the block boundary. The number of edge pixels is calculated by applying Sobel edge filter. An appropriate filtering operation is performed for each mode in both the horizontal and vertical directions. First, when the mode is associated with a smooth region, a strong filtering operation is applied because flat region are more sensitive to human visual system. In the second mode, adaptive filter that is based on the frequency information around the block boundaries is applied. Artifacts are reduced without introducing undesired distortion while preserving the original edge of image. Although the proposed approach is simple and operates on spatial domain, experimental results show that it improves both the subjective and objective quality of coded images with various features.

Real-time face detection and phone-to-face distance measuring for speech recognition for multi-modal interface in mobile device

This paper presents a real-time face detection and phone-to-face distance measurement for speech recognition in smart phones. Facial features such as sizes, shapes, positions, skin colors, and tilts as well as the surrounding environment, including lighting and complex background increases the difficulty of accurate detection of the speaker’s face. The proposed algorithm takes advantage of less-sensitive colors for face detection to overcome these problems. First, it converts the input RGB input into YCbCr image, and uses the magnitude value and Cb to Cr ratio color information in the YCbCr component. The algorithm extracts two candidate face regions based on the Cb to Cr ratio and the data obtained from experiments. It detects the final face region using vertical and horizontal histogram after integrating these two images into one. Experimental results demonstrate that the proposed algorithm has a higher detection rate than the existing methods with minimum complexity, facilitating real-time face detection.

Adaptive directional wavelet transform based on edge directed interpolation for image coding

We propose a new image compression method based on directional wavelet transform. The proposed method exploits the local characteristics of image by combining the wavelet filters adaptive to edge orientation with quincunx sub sampling to form a 2-D non separable transform through lifting. An edge directed interpolation is implemented to obtain the pixels in sub pixel location for better representation of edge. Experimental results show that the better performance can be achieved from the proposed algorithm in terms of bit rate and the subjective quality compared to image coding based on conventional separable wavelet transform.