Random access for remote visualization of an arbitrary 2-D planar cut from a large volumetric dataset poses significant challenges in terms bandwidth, storage and memory requirements. In our previous work, we have proposed a fast random image 3D retrieval system, which use overlapping rotated tiles to represent the dataset. This approach achieves a significant reduction in average transmission rate as compared to traditional cubic tiling. Additionally, this approach leads to improved random access, with less storage and runtime memory required at the client. In this paper, we first describe the details of our 3D compression technique. Then, we propose a new wavelet based approach, which can achieve 25% reduction in the average transmission rate at the cost of a factor of five in storage overhead (half of what was required for our previously proposed method).