International Conference Papers
Permanent URI for this collectionhttp://dl.cerist.dz/handle/CERIST/4
Browse
2 results
Search Results
Item H.264/AVC Data Hiding Based on Intra Prediction Modes for Real-time Applications(2012-10-24) Bouchama, Samira; Hamami, Latifa; Aliane, HassinaThe existing data hiding methods for the newest video codec H.264/AVC exploit its several modules such as the discrete cosine transform coefficients or the prediction modes. In this paper, a new data hiding approach is presented by exploiting the intra prediction modes for the 4x4 luminance blocks. The objective is to ensure a relatively high embedding capacity and to preserve the encoding and the decoding times in order to satisfy real-time applications. The intra prediction modes are divided into four groups composed of modes of close prediction directions. The data embedding is based on modifying modes of the same group in order to maintain visual quality and limit the number of additional calculation procedures. The increase of embedding capacity relies on the group composed on four modes since it allows the embedding of two bits per mode.Item H.264/AVC Data Hiding Algorithm with a Limitation of the Drift Distortion(2012-10-21) Bouchama, Samira; Hamami, Latifa; Aliane, HassinaMany data hiding algorithms have been proposed for the latest video codec H.264/AVC, most of them are based on the 4x4 luma DCT coefficients. However, the drift distortion is the main reason which limits the embedding capacity in data hiding algorithms based on DCT coefficients. Few methods have been proposed to compensate or eliminate the error propagation. Though, they are either non-blind, only detectable or need prior knowledge of the encoded blocks and thus cannot be used for real time broadcasting. In this paper we show that it is possible to reduce considerably the error propagation for real time applications. The proposed algorithm exploits the method of watermarking paired-coefficients in each block in order to bring the error to the middle of the block matrix. We evaluate the distortion caused by each paired-coefficients in order to give a watermarking priority to the pairs which introduce the minimum error. The proposed scheme offers a very good compromise between the video distortion, the increase in bitrate and the embedding capacity