Fuzzy set theoretic measure for automatic feature evaluation
IEEE Transactions on Systems, Man and Cybernetics
Representation of Uncertainty in Computer Vision Using Fuzzy Sets
IEEE Transactions on Computers
Image enhancement and thresholding by optimization of fuzzy compactness
Pattern Recognition Letters
Unsupervised Optimal Fuzzy Clustering
IEEE Transactions on Pattern Analysis and Machine Intelligence
Example-Based Learning for View-Based Human Face Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Learning to Parse Pictures of People
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part IV
A decision-theoretic generalization of on-line learning and an application to boosting
EuroCOLT '95 Proceedings of the Second European Conference on Computational Learning Theory
Detecting Pedestrians Using Patterns of Motion and Appearance
ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
Fuzzy component based object detection
International Journal of Approximate Reasoning
| Hi-index | 0.00 |
We present a human-centric framework for pattern classification. We call the framework human-centric because the classifier depends on the judgment and prior experience of a human expert for the interpretation of weak learner scores. In the first step a large number of simple Fuzzy Inference Engines (FIEs) are constructed to perform classification based on linguistic rules for weak learner score interpretation. The linguistic rules are simple if-then-else type conditions imposed on the weak learner scores combined with various membership functions and logical AND-OR-NOT type operators. A large number of FIEs are automatically generated by modifying the type or parameters of each membership function. The AdaBoost algorithm is used to find a reduced set of Fuzzy engines from a pool of FIEs. The detection rate and false positive rate on face detection data have been found to be comparable to other popular face detection algorithms. The processing time for each pattern is constrained only by the time taken by the input weak learner to come-up with a score. The FIEs always takes the same amount of processing time irrespective of the size of the image.