Name
Abstract | ||
|---|---|---|
Labeling faces in news video with their names is an interesting research problem which was previously solved using supervised methods that demand significant user efforts on labeling training data. In this paper, we investigate a more challenging setting of the problem where there is no complete information on data labels. Specifically, by exploiting the uniqueness of a face's name, we formulate the problem as a special multi-instance learning (MIL) problem, namely exclusive MIL or eMIL problem, so that it can be tackled by a model trained with partial labeling information as the anonymity judgment of faces, which requires less user effort to collect. We propose two discriminative probabilistic learning methods named Exclusive Density (ED) and Iterative ED for eMIL problems. Experiments on the face labeling problem shows that the performance of the proposed approaches are superior to the traditional MIL algorithms and close to the performance achieved by supervised methods trained with complete data labels. |
Year | DOI | Venue |
|---|---|---|
2005 | 10.1145/1101149.1101155 | ACM Multimedia 2001 |
Keywords | Field | DocType |
data label,training data,exclusive mil,complete information,supervised method,broadcasting news video,complete data label,interesting research problem,iterative ed,multiple instance,emil problem,traditional mil algorithm,machine learning | Training set,Broadcasting,Computer science,Artificial intelligence,Probabilistic logic,Anonymity,Discriminative model,Machine learning,Complete information,Labeling Problem | Conference |
ISBN | Citations | PageRank |
1-59593-044-2 | 41 | 1.60 |
References | Authors | |
21 | 3 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jun Yang | 1 | 937 | 37.42 |
| Rong Yan | 2 | 2019 | 104.99 |
| Alexander G. Hauptmann | 3 | 7472 | 558.23 |