Multiresolution Hash Encoding

Computer Graphics · Neural Representations · 2022 · multiresolution-hash-encoding.yaml

Replaces most of the network's capacity with trainable feature vectors in hash tables at several resolutions, so a tiny MLP suffices and collisions are resolved by the optimizer rather than avoided.

• supersedescorrects · extends
• classifiesspecializes · part-of
• substitutes forapproximates · alternative-to
• depends onrequires · validates

Colour is the family; a dashed line is the second member of it.

This node

• correctsfixes a defect in Neural Radiance Fieldsthe deep MLP, not the idea, is what made it slow

Referenced by

• requiresInstant NGP does not work without thisthe speedup is the encoding, not the network

References

• Instant Neural Graphics Primitives with a Multiresolution Hash Encoding — Muller, Evans, Schied, Keller — 2022 · link