Presentation is loading. Please wait.

Presentation is loading. Please wait.

Contributions A people dataset of 8035 images. Three layer attribute classification framework using poselets. 1 2.

Similar presentations


Presentation on theme: "Contributions A people dataset of 8035 images. Three layer attribute classification framework using poselets. 1 2."— Presentation transcript:

1

2 Contributions A people dataset of 8035 images. Three layer attribute classification framework using poselets. 1 2

3 People Dataset H3D (Humans in 3D)Pascal VOC 2010 (trn+val) 8035 images TRN: 2003VAL: 2011TEST: 4022 9 different attributes in total. At least 2 attributes per image Agreement of 4 out of 5 1

4 Attribute classification using poselets 2

5 What is a Poselet ? Poselets capture part of the pose from a given viewpoint [Bourdev & Malik, ICCV09]

6 Poselets Examples may differ visually but have common semantics [Bourdev & Malik, ICCV09]

7 Poselets But how are we going to create training examples of poselets?

8 How do we train a poselet for a given pose configuration?

9 Finding correspondences at training time Given part of a human pose How do we find a similar pose configuration in the training set?

10 We use keypoints to annotate the joints, eyes, nose, etc. of people Left Hip Left Shoulder Finding correspondences at training time

11 Residual Error Finding correspondences at training time

12 Training poselet classifiers Residual Error: 0.150.200.100.350.150.85 1. Given a seed patch 2. Find the closest patch for every other person 3. Sort them by residual error 4. Threshold them

13 Training poselet classifiers 1. Given a seed patch 2. Find the closest patch for every other person 3. Sort them by residual error 4. Threshold them 5. Use them as positive training examples to train a linear SVM with HOG features

14 Which poselets should we train? Choose thousands of random windows, generate poselet candidates, train linear SVMs Select a small set of poselets that are: –Individually effective –Complementary

15 Selecting a small set of complementary poselets

16 Some Poselets

17 Attribute classification using poselets 2

18 Features HOGs at two levels (~2K-4K features) –16 x 16 –32 x 32 Color Histograms in H,S,B (30 features) –10 bins for H, S and B Skin classifier output (3 features) –GMM with 5 components –Fraction of skin pixels –hands-skin, legs-skin, neck skin

19 Poselet-level Attribute Classifiers

20 Person-level Attribute Classifiers

21 Context-level Attribute Classifiers

22 Results

23

24

25

26 Gender Classification Results

27 Thanks


Download ppt "Contributions A people dataset of 8035 images. Three layer attribute classification framework using poselets. 1 2."

Similar presentations


Ads by Google