Zh-Transform and Its Application in Video Object Shape Coding

In order to improve the scalability of bitmap-based shape coding methods, a new transform called Zh-transform is presented to deal with binary bitmap. After Zh-transform, the original bitmap representing the shape of video object is decomposed into several small bitmaps. These small bitmaps describe the shape from rough glancing to fine detailing. Based on Zh-transform and context-based arithmetic encoding, a new shape coding method is formulated. Experiment results show that the new method not only has good spatial and SNR scalability but also has much higher coding efficiency than the MPEG-4 first edition.