256 search results - page 3 / 52 » Multiple-instance ranking: Learning to rank images for image... |
NIPS
2007
13 years 7 months ago
2007
We present a framework for active learning in the multiple-instance (MI) setting. In an MI learning problem, instances are naturally organized into bags and it is the bags, instea...
FLAIRS
2004
13 years 7 months ago
2004
In our prior work, we introduced a generalization of the multiple-instance learning (MIL) model in which a bag's label is not based on a single instance's proximity to a...
ISBI
2009
IEEE
14 years 1 months ago
2009
IEEE
One of the main challenges for Content-Based Image Retrieval (CBIR) is to achieve meaningful mappings between the high-level semantic concepts and the low-level visual features in...
DIS
2009
Springer
14 years 24 days ago
2009
Springer
Multiple-instance learning (MIL) is a generalization of the supervised learning problem where each training observation is a labeled bag of unlabeled instances. Several supervised ...
MIR
2010
ACM
13 years 4 months ago
2010
ACM
In Content-based Image Retrieval (CBIR), accurately ranking the returned images is of paramount importance, since users consider mostly the topmost results. The typical ranking st...