ComfyUI Node: Preview 360 Panorama

Preview 360 Panorama:

The PanoramaViewerNode is a specialized node within the ComfyUI framework designed to provide an interactive 360-degree panorama viewing experience for equirectangular images. This node leverages a Three.js-based viewer to allow you to explore panoramic images with ease. You can pan around the 360-degree view using mouse drag and zoom in or out using the mouse wheel, all while maintaining the correct equirectangular projection. The node is adept at handling various image formats and sizes, offering features such as batch processing by selecting the first image, converting grayscale images to RGB, resizing images for optimal performance, and converting float image data to uint8 format. This makes it an invaluable tool for artists and designers who wish to preview and interact with panoramic images in a seamless and efficient manner.

Preview 360 Panorama Input Parameters:

images

The images parameter is a torch.Tensor that contains the panoramic image or images you wish to view. It can be in the format of either (B, H, W, C) for batches or (H, W, C) for single images. This parameter is crucial as it provides the raw image data that the node processes and displays in the panoramic viewer. The node will automatically handle batch dimensions by selecting the first image, and it will convert the tensor to a numpy array for further processing.

max_width

The max_width parameter is an integer that specifies the maximum width for resizing the image. The default value is 4096, meaning images larger than this width will be resized to optimize performance. If you set this parameter to -1, resizing will be disabled, allowing you to view images in their original size. This parameter is important for managing the performance and quality of the panoramic view, especially when dealing with high-resolution images.

Preview 360 Panorama Output Parameters:

ui

The ui output parameter is a dictionary containing the base64-encoded PNG image data, which is used to update the user interface with the processed panoramic image. This output is essential as it provides the necessary data for rendering the image in the Three.js-based viewer, allowing you to interact with the panorama seamlessly. The base64 encoding ensures that the image can be easily embedded and displayed in web-based interfaces.

Preview 360 Panorama Usage Tips:

• To ensure optimal performance, consider resizing large images by setting an appropriate max_width value. This can help reduce the computational load and improve the responsiveness of the viewer.
• When working with batches of images, remember that the node will only process the first image in the batch. If you need to view multiple images, process them individually.
• If your images are in grayscale, the node will automatically convert them to RGB, so you don't need to preprocess them manually.

Preview 360 Panorama Common Errors and Solutions:

Image tensor shape mismatch

• Explanation: This error occurs when the input tensor does not conform to the expected shape of (B, H, W, C) or (H, W, C).
• Solution: Ensure that your input tensor is correctly shaped. If you are working with a single image, it should be in the format (H, W, C). For batches, use (B, H, W, C).

Unsupported image data type

• Explanation: The node expects the image data to be in a format that can be converted to uint8. - Solution: Verify that your image data is in a compatible format, such as float values that can be scaled to uint8. If necessary, preprocess your images to ensure compatibility.

Image exceeds maximum width

• Explanation: The input image is larger than the specified max_width, which may lead to performance issues.
• Solution: Consider resizing your image before inputting it into the node, or adjust the max_width parameter to accommodate larger images.