Layered depth images (LDIs) are a representation that allows efficient image-based rendering. The paper presents two new methods for image-based rendering: Sprites with Depth and Layered Depth Images. Sprites with Depth are 2D images with depth information that can be warped to simulate parallax. Layered Depth Images are a more general representation that can handle complex geometries and large parallax effects. Both methods allow for efficient rendering of multiple frames per second on a PC. The paper discusses the limitations of traditional texture mapping and how Sprites with Depth and Layered Depth Images address these issues. Sprites with Depth use depth information to add parallax corrections to a standard sprite renderer. Layered Depth Images contain multiple depth pixels at each discrete location in the image, allowing for more accurate representation of complex scenes. The paper also describes the rendering process for Sprites with Depth and Layered Depth Images. Sprites with Depth use a two-step rendering algorithm that first forward maps the displacement map and then performs a backward mapping step. Layered Depth Images use a similar approach, but with the additional benefit of being able to handle more complex geometries. The paper also discusses the use of layered depth images in computer vision applications, where they can be extracted from image sequences using a layered motion estimation algorithm. This algorithm segments the sequence into coherently moving regions and computes a parametric motion estimate for each layer. The paper concludes with results showing the effectiveness of the proposed methods. Sprites with Depth and Layered Depth Images allow for efficient rendering of multiple frames per second on a PC, and can be used in a variety of applications, including image-based rendering and computer vision.Layered depth images (LDIs) are a representation that allows efficient image-based rendering. The paper presents two new methods for image-based rendering: Sprites with Depth and Layered Depth Images. Sprites with Depth are 2D images with depth information that can be warped to simulate parallax. Layered Depth Images are a more general representation that can handle complex geometries and large parallax effects. Both methods allow for efficient rendering of multiple frames per second on a PC. The paper discusses the limitations of traditional texture mapping and how Sprites with Depth and Layered Depth Images address these issues. Sprites with Depth use depth information to add parallax corrections to a standard sprite renderer. Layered Depth Images contain multiple depth pixels at each discrete location in the image, allowing for more accurate representation of complex scenes. The paper also describes the rendering process for Sprites with Depth and Layered Depth Images. Sprites with Depth use a two-step rendering algorithm that first forward maps the displacement map and then performs a backward mapping step. Layered Depth Images use a similar approach, but with the additional benefit of being able to handle more complex geometries. The paper also discusses the use of layered depth images in computer vision applications, where they can be extracted from image sequences using a layered motion estimation algorithm. This algorithm segments the sequence into coherently moving regions and computes a parametric motion estimate for each layer. The paper concludes with results showing the effectiveness of the proposed methods. Sprites with Depth and Layered Depth Images allow for efficient rendering of multiple frames per second on a PC, and can be used in a variety of applications, including image-based rendering and computer vision.