Skip to main content
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub
This node removes duplicate or very similar images from a batch. It works by creating a perceptual hash for each image—a simple numerical fingerprint based on its visual content—and then comparing them. Images whose hashes are more similar than a set threshold are considered duplicates and filtered out.

Inputs

ParameterData TypeRequiredRangeDescription
imagesIMAGEYes-The batch of images to process for deduplication.
similarity_thresholdFLOATNo0.0 - 1.0Similarity threshold (0-1). Higher means more similar. Images above this threshold are considered duplicates. (default: 0.95)

Outputs

Output NameData TypeDescription
imagesIMAGEThe filtered list of images with duplicates removed.