集成学习课件.pptVIP

Scanning the Detector Scaling is achieved by scaling the detector itself, rather than scaling the image Good detection results for scaling factor of 1.25 The detector is scanned across location Subsequent locations are obtained by shifting the window [s?] pixels, where s is the current scale The result for ? = 1.0 and ? = 1.5 were reported Merging Multiple Detections ROC Curves for Face Detection Output of Face Detector on Test Images Other Detection Tasks Facial Feature Localization Male vs. female Profile Detection Other Detection Tasks Facial Feature Localization Profile Detection Other Detection Tasks Male vs. Female Conclusions How adaboost works? Why adaboost works? Adaboost for face detection Rectangle features Integral images for fast computation Boosting for feature selection Attentional cascade for fast rejection of negative windows * * * * 这里区分学习器和分类器的区别。学习器泛指任意的学习问题，分类器指分类问题（一般限制在两类） * Ensemble is a method like Majority vote Suppose we have 5 completely independent classifiers…If accuracy is 70% for each, we can get 83.7% accuracy by majority vote. If we have 101 such classifiers then we can get 99.9% majority vote accuracy * 如果来来去去都是覆盖同一块，就没有意思了。 * AdaBoost Its Applications AdaBoost for Face Detection The Task ofFace Detection Many slides adapted from P. Viola Basic Idea Slide a window across image and evaluate a face model at every location. Image Features Rectangle filters Image Features Rectangle filters Size of Feature Space How many number of possible rectangle features for a 24x24 detection region? Rectangle filters A+B C D Feature Selection How many number of possible rectangle features for a 24x24 detection region? A+B C D What features are good for face detection? Feature Selection How many number of possible rectangle features for a 24x24 detection region? A+B C D Can we create a good classifier using just a small subset of all possible features? How to select such a subset? Integral images The integral image computes a value at each