Examples
Examples
Try the starting example
- Images: unique image ID, bound SKUs joined by |, equal-length hexadecimal hash
- I1,A,00 I2,B,01 I3,C,03 I4,D,ff
- Maximum pairwise Hamming distance for a graph edge
- 1
- Connected candidate image clusters
- I1 | I2 | I3
- Candidate cluster count
- 1
- Images beyond one representative per cluster
- 2
- Largest within-cluster pair distances
- 2
- Clusters spanning multiple bound SKUs
- I1 | I2 | I3
Try another scenario
- Images: unique image ID, bound SKUs joined by |, equal-length hexadecimal hash
- I1,A,00 I2,B,01 I3,C,03 I4,D,ff
- Maximum pairwise Hamming distance for a graph edge
- 0
- Connected candidate image clusters
- None
- Candidate cluster count
- 0
- Images beyond one representative per cluster
- 0
- Largest within-cluster pair distances
- None
- Clusters spanning multiple bound SKUs
- None
Calculation method
Calculation method
Connected candidate image clusters
Join image nodes when supplied hashes differ by at most the threshold; report connected components larger than one.
Candidate cluster count
Count connected components with two or more images.
Images beyond one representative per cluster
Sum(cluster size−1), not a deletion recommendation.
Largest within-cluster pair distances
Report maximum actual pair distance within each connected cluster, which can exceed the edge threshold through transitive links.
Clusters spanning multiple bound SKUs
Report candidate clusters whose combined binding set contains multiple SKUs.
Symbols and units
- a
- Images: unique image ID, bound SKUs joined by |, equal-length hexadecimal hash
- b
- Maximum pairwise Hamming distance for a graph edge
Use the units shown. A list keeps the order in which you enter its values. The symbol ^ means “raised to the power of”.
Assumptions & limits
Assumptions & limits
- A supplied-hash graph audit, without image generation, visual inspection or file deletion. Hash closeness and transitive membership are candidate signals, not proof of identical products or images.
Are these inputs private?
These calculations run in your browser. A shared link contains the values you choose to share. Only questions submitted to the separate AI finder are sent to OpenAI.
Further reading
Google Merchant Center: product-data terminology and consistency background ↗
Background reading; this does not imply endorsement or independent review of SumFizz.