Evaluated Algorithms#
Info
The content of this section is autogenerated based on the latest published configuration of the TwinSpect Benchmark.
ISCC Text-Code V0 64-Bit#
The ISCC Text-Code is a similarity preserving hash designed to cluster and match near-duplicate text documents that have undergone format conversion or minor edits.
The reference implementation is available in the
iscc/iscc-core - iscc_core/code_content_text.py
GitHub Repository
ISCC Image-Code V0 64-Bit#
The ISCC Image-Code is a similarity preserving perceptual hash designed to cluster and match near-duplicate images that have undergone format conversion or minor edits.
The reference implementation is available in the
iscc/iscc-core - iscc_core/code_content_image.py
GitHub Repository
ISCC Audio-Code V0 64-Bit#
The ISCC Audio-Code is a similarity preserving hash based on chromaprint and designed to cluster and match near-duplicate audio files that have undergone format conversion, transcoding, compression and other minor edits.
The reference implementation is available in the
iscc/iscc-core - iscc_core/code_content_audio.py
GitHub Repository
ISCC Video-Code V0 64-Bit#
The ISCC Video-Code is a similarity preserving hash based on the MPEG-7 Video Signature and is designed to cluster and match near-duplicate videos that have undergone format conversion or minor edits.
The reference implementation is available in the
iscc/iscc-core - iscc_core/code_content_video.py
GitHub Repository
ISCC Image-Code-S 64-Bit#
The ISCC Image-Code-S is a semantic image similarity hash based on a deep neural network (DINOv2). It is designed to cluster and match semantically similar images independent of visual transformations.
The reference implementation is available in the iscc/iscc-sci
GitHub Repository
ISCC Image-Code-SC 128-Bit#
Combined 128-bit image code concatenating the 64-bit semantic code (iscc-sci/DINOv2) with the 64-bit perceptual code (iscc-sdk). Captures both semantic similarity and visual characteristics.
The reference implementation is available in the
iscc/twinspect GitHub Repository
ISCC Text-Code-S 64-Bit#
The ISCC Text-Code-S is a semantic text similarity hash based on a deep neural network (DeBERTa). It is designed to cluster and match semantically similar text independent of textual transformations, supporting cross-lingual similarity matching.
The reference implementation is available in the iscc/iscc-sct
GitHub Repository
ISCC Text-Code-SC 128-Bit#
Combined 128-bit text code concatenating the 64-bit semantic code (iscc-sct/DeBERTa) with the 64-bit perceptual code (iscc-sdk). Captures both semantic similarity and textual characteristics.
The reference implementation is available in the
iscc/twinspect GitHub Repository