1. An Exemplar Model for Learning Object Classes Authors: Ondrej Chum Andrew Zisserman@University of Oxford Presenter: Shao-Chuan Wang
2. An Exemplar Model for Learning Object Classes Objective: Give training images known to contain instances of an object class, without specifying locations and scales. Detect and localize object Kea Ideas: Learn region of interest (ROI) around class instance in weakly supervised training data. Based on discriminative features to initialize ROI for the optimization problem
3. An Exemplar Model for Learning Object Classes Exemplar model: Detection (cost function): X Y X: exemplar set X^w: PHOW descriptor X^e: PHOG descriptor A: aspect ratio of target region d: distance function /mu: mean of exemplars’ aspect ratio /sigma: std of exemplars’ aspect ratio /alpha, /beta: weighting to be tuned/learned
4. An Exemplar Model for Learning Object Classes Learning the exemplar model: Learn the regions in all images simultaneously. How to Determine initial ROI? > By discriminative features
5. Top 10 most discriminative visual words Discriminative features Definition:
7. Constructing ROI exemplars: Algorithm Initialization Calculate discriminability of visual words Initialize the ROI in each training image by a bounding box of the 64 most discriminative features Optimization of cost function Find the ROI to minimize the cost function with eta = 0 Re-initialization by detection Refinement Enlarge the ROI in the training images by 10% Calculate discriminability of visual words using only the features inside the ROI Optimization of cost function (goto 2.)
8. Constructing ROI exemplars: Algorithm Initialization Calculate discriminability of visual words Initialize the ROI in each training image by a bounding box of the 64 most discriminative features Optimization of cost function Find the ROI to minimize the cost function with eta = 0 Re-initialization by detection Refinement Enlarge the ROI in the training images by 10% Calculate discriminability of visual words using only the features inside the ROI Optimization of cost function (goto 2.)
9. Constructing ROI exemplars: Algorithm Initialization Calculate discriminability of visual words Initialize the ROI in each training image by a bounding box of the 64 most discriminative features Optimization of cost function Find the ROI to minimize the cost function with eta = 0 Re-initialization by detection Refinement Enlarge the ROI in the training images by 10% Calculate discriminability of visual words using only the features inside the ROI Optimization of cost function (goto 2.)
10. Constructing ROI exemplars: Algorithm Initialization Calculate discriminability of visual words Initialize the ROI in each training image by a bounding box of the 64 most discriminative features Optimization of cost function Find the ROI to minimize the cost function with eta = 0 Re-initialization by detection. Refinement Enlarge the ROI in the training images by 10% Calculate discriminability of visual words using only the features inside the ROI Optimization of cost function (goto 2.)
11. Constructing ROI exemplars: Algorithm Initialization Calculate discriminability of visual words Initialize the ROI in each training image by a bounding box of the 64 most discriminative features Optimization of cost function Find the ROI to minimize the cost function with eta = 0 Re-initialization by detection. Refinement Enlarge the ROI in the training images by 10% Calculate discriminability of visual words using only the features inside the ROI Optimization of cost function (goto 2.)
12. Constructing ROI exemplars: Algorithm Initialization Calculate discriminability of visual words Initialize the ROI in each training image by a bounding box of the 64 most discriminative features Optimization of cost function Find the ROI to minimize the cost function with eta = 0 Re-initialization by detection. Refinement Enlarge the ROI in the training images by 10% Calculate discriminability of visual words using only the features inside the ROI Optimization of cost function (goto 2.)
13. Constructing ROI exemplars: Algorithm Three stages of the optimization process Initialization Optimization Re-initialization via detection
14. Using the exemplar model Object Detection Hypothesis Score of a hypothesis n_(w,R): the number of exemplar Images consistent with the hypothesis #w: the number of appearances of the visual word w in the exemplar images Clustering 20 strongest hypotheses are tested on each test image
15. Using other models Training: Train an SVM, using features within ROI by exemplar models Object detection Scores are ranked by SVM score
17. Conclusion When constructing exemplars’ ROI, they use discriminability to initialize bounding box In detection, they used relative position of bounding boxes and visual words to try the most probable hypotheses. It may failed to detect when significant class variability in the exemplars, such as people class.