ComfyUI Node: ◎ Radiance Depth Map

◎ Radiance Depth Map:

The RadianceDepthMapGenerator is a sophisticated node designed to provide high-quality depth estimation for images, leveraging the advanced capabilities of the Depth Anything V2 model. This node is particularly beneficial for AI artists and creators who require precise depth maps to enhance their visual projects. It automatically downloads the necessary model from HuggingFace, ensuring ease of use and accessibility. With an impressive accuracy rate of 97.1% and a rapid inference time of 213 milliseconds, this node is optimized for both speed and quality. It is capable of handling complex surfaces, including transparent and reflective ones, making it a versatile tool for generating grayscale depth maps where white represents near objects and black represents far ones. This depth information can be seamlessly integrated with other nodes, such as the Depth of Field node, to create realistic blur effects in images.

◎ Radiance Depth Map Input Parameters:

image

The image parameter is a tensor representing the input image from which the depth map will be generated. This parameter is crucial as it serves as the primary data source for the depth estimation process. The quality and characteristics of the input image directly influence the accuracy and detail of the resulting depth map.

model_size

The model_size parameter determines the size of the depth estimation model to be used. It offers three options: "Small (25M

• Fast)" for quick previews, "Base (98M - Balanced)" for a balance between speed and quality, and "Large (335M - Best)" for the highest quality suitable for final renders. The choice of model size impacts the processing time and the quality of the depth map.

normalize

The normalize parameter is a boolean that, when set to true, normalizes the depth map values to a standard range. This helps in maintaining consistency across different images and ensures that the depth map is easily interpretable. The default value is true.

invert

The invert parameter is a boolean that, when enabled, inverts the depth map, swapping the representation of near and far objects. This can be useful for specific artistic effects or when integrating with other systems that require inverted depth data. The default value is false.

blur_edges

The blur_edges parameter is a float that controls the amount of edge blurring applied to the depth map. It ranges from 0.0 to 5.0, with a default value of 0.0. This parameter is useful for smoothing out transitions in the depth map, which can enhance the visual quality of the final output.

use_gpu

The use_gpu parameter is a boolean that specifies whether to utilize GPU acceleration for processing. Enabling this option can significantly speed up the depth estimation process, especially for larger models and images. The default value is true.

◎ Radiance Depth Map Output Parameters:

depth_map

The depth_map is the primary output of the node, represented as an image. It is a grayscale representation where white indicates objects that are closer to the camera, and black indicates objects that are further away. This output is essential for creating depth-based effects in images, such as realistic blurring or 3D transformations.

◎ Radiance Depth Map Usage Tips:

• For quick previews, use the "Small (25M - Fast)" model size to save time while still getting a reasonable depth estimation.
• When working on final renders where quality is paramount, opt for the "Large (335M
• Best)" model size to ensure the highest fidelity in depth representation.
• Utilize the normalize parameter to maintain consistent depth map outputs across different images, which is particularly useful when batch processing.
• If you encounter performance issues, ensure that the use_gpu parameter is enabled to leverage GPU acceleration for faster processing.

◎ Radiance Depth Map Common Errors and Solutions:

transformers library not installed

• Explanation: This error occurs when the required transformers library is not installed in your environment.
• Solution: Install the transformers library by running the command pip install transformers in your terminal or command prompt.

Failed to download depth model

• Explanation: This error indicates that the node was unable to download the depth model from HuggingFace, possibly due to network issues or incorrect model identifiers.
• Solution: Check your internet connection and ensure that the model identifiers are correct. Retry the operation after verifying these aspects.