ScholarGate
Assistent
Machine learningDeep learning / NLP / CV

Forklarlig instanssegmentering

Forklarlig instanssegmentering kombinerer deep learning-modeller til instanssegmentering – som detekterer og afgrænser hvert enkelt objekt som en separat pixelmaske – med post-hoc eller ante-hoc forklarbarhedsteknikker som GradCAM, SHAP, LIME eller opmærksomhedsvisualisering, så hver forudsagt maske ledsages af evidens, der viser, hvilke billedregioner der drev modellens beslutning.

Åbn i MethodMindSnartVideoSnartDownload slides

Læs hele metoden

Kun for medlemmer

Log ind med en gratis konto for at læse dette afsnit.

Log ind

Method map

The neighbourhood of related methods — select a node to explore.

Kilder

  1. Lindner, M., Meng, C., & Bischl, B. (2023). Explaining Instance Segmentation Models via Saliency Maps and Occlusion. IEEE Transactions on Pattern Analysis and Machine Intelligence. link
  2. Instance segmentation. Wikipedia. link

Sådan citerer du denne side

ScholarGate. (2026, June 3). Explainable Instance Segmentation (XAI-augmented Mask Detection). ScholarGate. https://scholargate.app/da/deep-learning/explainable-instance-segmentation

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

Compare side by side
ScholarGateExplainable Instance Segmentation (Explainable Instance Segmentation (XAI-augmented Mask Detection)). Hentet 2026-06-15 fra https://scholargate.app/da/deep-learning/explainable-instance-segmentation · Datasæt: https://doi.org/10.5281/zenodo.20539026