Name
Abstract | ||
|---|---|---|
Many object detectors, such as AdaBoost, SVM and deformable part-based models (DPM), compute additive scoring functions at a large number of windows scanned over image pyramid, thus computational efficiency is an important consideration beside accuracy performance. In this paper, we present a framework of learning cost-sensitive decision policy which is a sequence of two-sided thresholds to execute early rejection or early acceptance based on the accumulative scores at each step. A decision policy is said to be optimal if it minimizes an empirical global risk function that sums over the loss of false negatives (FN) and false positives (FP), and the cost of computation. While the risk function is very complex due to high-order connections among the two-sided thresholds, we find its upper bound can be optimized by dynamic programming (DP) efficiently and thus say the learned policy is near-optimal. Given the loss of FN and FP and the cost in three numbers, our method can produce a policy on-the-fly for Adaboost, SVM and DPM. In experiments, we show that our decision policy outperforms state-of-the-art cascade methods significantly in terms of speed with similar accuracy performance. |
Year | DOI | Venue |
|---|---|---|
2013 | 10.1109/ICCV.2013.98 | Pattern Analysis and Machine Intelligence, IEEE Transactions |
Keywords | DocType | Volume |
false negative,cost-sensitive computing,early acceptance,image pyramid,decision policy,two-sided threshold,policy on-the-fly,deformable part-based models,risk minimization,learning (artificial intelligence),computational efficiency,object detection,accuracy performance,dynamic programming,cost-sensitive decision policy,svm,adaboost,offalse negatives,additive scoring functions,learning near-optimal cost-sensitive decision policy,object detectors,empirical global risk function,additive scoring function,learning near-optimal cost-sensitive decision,early rejection,two-sided thresholds,false positives,learning artificial intelligence | Conference | 37 |
Issue | ISSN | Citations |
5 | 1550-5499 | 5 |
PageRank | References | Authors |
0.42 | 36 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Tianfu Wu | 1 | 331 | 26.72 |
| Song-Chun Zhu | 2 | 6580 | 741.75 |