Presentation is loading. Please wait.

Presentation is loading. Please wait.

Evaluating Color Descriptors for Object and Scene Recognition Koen E.A. van de Sande, Student Member, IEEE, Theo Gevers, Member, IEEE, and Cees G.M. Snoek,

Similar presentations


Presentation on theme: "Evaluating Color Descriptors for Object and Scene Recognition Koen E.A. van de Sande, Student Member, IEEE, Theo Gevers, Member, IEEE, and Cees G.M. Snoek,"— Presentation transcript:

1 Evaluating Color Descriptors for Object and Scene Recognition Koen E.A. van de Sande, Student Member, IEEE, Theo Gevers, Member, IEEE, and Cees G.M. Snoek, Member, IEEE IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 32, NO. 9, SEPTEMBER 2010 1

2 Outline Introduction Reflectance Model Diagonal Model COLOR DESCRIPTORS AND INVARIANTPROPERTIES EXPERIMENTAL SETUP RESULTS Discussion Conclusion 2

3 Introduction To increase illumination invariance and discriminative power, color descriptors have been proposed. This paper studies the invariance properties and the distinctiveness of color descriptors. The usefulness of invariance is category-specific. Recommendations are given on which color descriptors to use under which circumstances and data sets. 3

4 Reflectance Model 4 An image f can be modeled under the assumption of Lambertian reflectance as follows: Shafer [23] proposes adding a diffuse term :

5 Reflectance Model The spatial derivative of f at location x on scale is given by: Hence, derivatives will yield invariance to diffuse light. 5

6 Diagonal Model 6 Changes in the illumination can be modeled by a diagonal mapping or von Kries Model [18] as follows: Diagonal-offset model:

7 Diagonal Model 1. light intensity changes 2. light intensity shifts 7

8 3. light intensity changes and shift 4. light color changes – Allowing (a ≠ b ≠ c) 5. light color changes and shift – (a ≠ b ≠ c) & (o1≠o2≠o3) 8

9 COLOR DESCRIPTORS AND INVARIANTPROPERTIES Three types of descriptors are discussed. – Histogram  do not contain local spatial information and are inherently pixel-based. – color moment and color moment invariant  contain local photometrical and spatial information derived from pixel values. – SIFT  contain local spatial information and are derivative-based. 9

10 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 10

11 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 11

12 COLOR DESCRIPTORS AND INVARIANTPROPERTIES Opponent histogram – Opponent color space: – O1 and O2 are shift-invariant with respect to light intensity. 12

13 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 13

14 COLOR DESCRIPTORS AND INVARIANTPROPERTIES Hue histogram – A certainty of the hue is inversely proportional to the saturation. – To be more robust, weigh each sample of the hue by its saturation. The H color model is scale-invariant and shift- invariant with respect to light intensity [14]. 14

15 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 15

16 COLOR DESCRIPTORS AND INVARIANTPROPERTIES rghistogram – The normalized RGB color model – r and g (b is redundant ) are scale-invariant with respect to light intensity changes, shadows, and shading. 16

17 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 17

18 COLOR DESCRIPTORS AND INVARIANTPROPERTIES Transformed color distribution – Scale-invariance and shift-invariance are achieved by normalizing the pixel value distributions. – Each channel is normalized independently, it also have invariance against changes in light color and arbitrary offsets. 18

19 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 19

20 COLOR DESCRIPTORS AND INVARIANTPROPERTIES SIFT – proposed by Lowe [9] describes the local shape of a region using edge orientation histograms. – The gradient of an image is shift-invariant – The descriptor is normalized, so it have scale- invariance. 20

21 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 21

22 COLOR DESCRIPTORS AND INVARIANTPROPERTIES HSV-SIFT – H color model is scale-invariant and shift-variant. – Complete descriptor have no invariance properties due to the combination of the HSV channels. HueSIFT – Similar to hue histogram, is scale-invariant and shift-invariant. 22

23 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 23

24 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 24

25 COLOR DESCRIPTORS AND INVARIANTPROPERTIES C-SIFT – C-invariant, which can be intuitively seen as the normalized opponent color space and. – C-SIFT is scale-invariant. 25

26 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 26

27 COLOR DESCRIPTORS AND INVARIANTPROPERTIES 27

28 COLOR DESCRIPTORS AND INVARIANTPROPERTIES RGB-SIFT – SIFT descriptors are computed for every RGB channel independently. – Its descriptor values are equal to the transformed color space. 28

29 EXPERIMENTAL SETUP Feature Extraction Pipelines 29

30 Experiment 1: Illumination Changes The Amsterdam Library of Object Images (ALOI) data set [20] contains more than 48,000 images of 1,000 objects, under various illumination conditions. – artificially added “light intensity scaling “ and “light intensity shifts “ to the data set. – The light color is varied by changing the illumination color temperature – Objects lighted by a different number of white lights at increasingly oblique angles – object rotation images – images with differentlevels of JPEG compression. 30

31 Experiment 2: Image Benchmark The PASCAL Visual Object Classes Challenge [21] provides a yearly benchmark for comparison of object classification systems. 31

32 Experiment 3: Video Benchmark The Mediamill Challenge by Snoek et al. [22] provides an annotated video data set, based on the training set of the NIST TRECVID 2005 benchmark [7]. 32

33 Evaluation Criteria The top-ranked result should be equal to the original image of the object for successful recognition. The percentage of objects where the top-ranked result is indeed the correct object is used as the performance on the ALOI data set. The average precision is taken as the performance metric for determining the accuracy of ranked category recognition results. 33

34 RESULTS – Experiment1 The theoretical invariance properties of color descriptors are validated. The SIFT and color SIFT descriptors perform much better than histogram-based descriptors Certain color descriptors are sensitive to compression artifacts, reducing their usefulness. The descriptors with the best overall performance are C-SIFT, rgSIFT, OpponentSIFT, and RGB-SIFT. The increased invariance comes at the price of reduced discriminative power. 34

35 35 RESULTS – Experiment1

36 36 RESULTS – Experiment1

37 37 RESULTS – Experiment1

38 38

39 RESULTS – Experiment1 39

40 RESULTS – Experiment1 40

41 RESULTS – Experiment2 The SIFT variants perform significantly better than color moments, moment invariants, and color histograms. The additional invariance makes the descriptor less discriminative for these object categories because a reduction in performance is observed. 41

42 RESULTS – Experiment2 42

43 RESULTS – Experiment2 43

44 RESULTS – Experiment3 SIFT and color SIFT variants perform significantly better than the other descriptors. OpponentSIFT perform better than C-SIFT and rgSIFT for these categories that occur under a wide range of light intensities. 44

45 RESULTS – Experiment3 45

46 Comparison with State-of-the-Art 46

47 Discussion The SIFT-based descriptors outperform the other on both two category recognition. large variations in lighting conditions occur frequently invariance to light intensity shifts is useful. From the results, it can be noticed that invariance to light color changes and shifts is domain-specific. 47

48 Discussion 48

49 Discussion 49

50 Conclusion These invariance properties were validated using a data set with known photometric changes. The addition of color descriptors over SIFT improves category recognition by 8 percent and 7 percent, respectively. Choosing a single descriptor and no prior knowledge about the data set and object and scene categories is available. 50

51 Conclusion 51


Download ppt "Evaluating Color Descriptors for Object and Scene Recognition Koen E.A. van de Sande, Student Member, IEEE, Theo Gevers, Member, IEEE, and Cees G.M. Snoek,"

Similar presentations


Ads by Google