Multimodal Image Classification
Multimodal image classification extends standard visual classification by incorporating additional modalities — such as text captions, audio, or structured metadata — alongside image features. Separate encoders process each modality, their representations are fused, and a joint classifier assigns the target label. Models such as CLIP demonstrate that image–text alignment enables zero-shot and few-shot image classification at scale.
Record di origine
Citazioni copiate testualmente dal record di origine del metodo. Non si inferisce alcuna verifica a livello di affermazione da esse.
- Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., ... & Sutskever, I. (2021). Learning transferable visual models from natural language supervision. Proceedings of the 38th International Conference on Machine Learning (ICML), PMLR 139, 8748–8763. · URL
- Ngiam, J., Khosla, A., Kim, M., Nam, J., Lee, H., & Ng, A. Y. (2011). Multimodal deep learning. Proceedings of the 28th International Conference on Machine Learning (ICML), 689–696. · URL
Affermazioni curate
Affermazioni persistite nel registro delle evidenze, ciascuna con la propria valutazione.
Questa vista non inventa una valutazione dell'affermazione quando il registro non ne ha.
Metodi correlati
Generato dal grafo dei metodi e mostrato come relazioni suggerite dalla macchina — nessuna affermazione di evidenza viene inferita.