Gyeong Ryeol Song

IGNORE: Information Gap-based False Negative Loss Rejection for Single Positive Multi-Label Learning

"Single Positive Multi-Label Learning (SPML) is a method for a scarcely annotated setting, in which each image is assigned only one positive label while the other labels remain unannotated. Most approaches for SPML assume unannotated labels as negatives (“Assumed Negative”, AN). However, with this a…