Presentation is loading. Please wait.

Presentation is loading. Please wait.

Video Surveillance: Legally Blind? Peter Kovesi School of Computer Science & Software Engineering The University of Western Australia.

Similar presentations


Presentation on theme: "Video Surveillance: Legally Blind? Peter Kovesi School of Computer Science & Software Engineering The University of Western Australia."— Presentation transcript:

1 Video Surveillance: Legally Blind? Peter Kovesi School of Computer Science & Software Engineering The University of Western Australia

2

3

4 Questions What image quality do we need for identification? How do you measure image quality? What is the image quality from a video camera? What is the effect on image quality when you: record to video tape? use image compression?

5 Humans are very bad at recognizing unfamiliar faces Kemp, Towell and Pike (1997) tested the value of having photos on credit cards. When a user presented a card with a photograph of someone else that had some resemblance to the user, they were challenged less than 40% of the time. Bruce et al. (1999, 2001) have tested the ability of people to match good quality CCTV images of unfamiliar faces under a variety of scenarios. Correct recognition rates are typically only 70-80%.

6 Good quality photograph of target Array of 10 good quality CCTV images Bruce et al (1999). Is this person in the array? If they are present match the person.

7 Good quality photograph of target Array of 10 good quality CCTV images Bruce et al (1999). Is this person in the array? If they are present match the person.

8 Good quality photograph of target Array of 10 good quality CCTV images When target was present in the array. 12% picked wrong person and 18% said they were not present (overall only 70% correct). When target was not present in the array 70% still matched the target to someone in the array. Bruce et al (1999). Is this person in the array? If they are present match the person.

9 Face recognition performance by humans is poor. Face recognition performance by machine is becoming quite good - but only if the images are of good quality. Surveillance video rarely provides good quality images.

10 Face recognition performance by humans is poor. Face recognition performance by machine is becoming quite good - but only if the images are of good quality. Surveillance video rarely provides good quality images. What image quality is needed for face identification?

11 Image quality is defined by many attributes Minimum feature size that can be resolved Noise level Quality of luminance reproduction Quality of colour reproduction.

12 (Hayes, Morrone and Burr 1986) (Costen, Parker and Craw 1996) (Nasanen 1999) In humans it has been found that face recognition is tuned to a set of spatial frequencies ranging from about 20 cycles per face width down to about 5 cycles per face width. 20 cycles 10 cycles 5 cycles Human Face Recognition Maximum sensitivity is centred around 8 to 13 cycles/face width. To recognize with confidence you need to be able to resolve down to 20 cycles/face width

13 (Hayes, Morrone and Burr 1986) (Costen, Parker and Craw 1996) (Nasanen 1999) In humans it has been found that face recognition is tuned to a set of spatial frequencies ranging from about 20 cycles per face width down to about 5 cycles per face width. ~ 160mm 20 cycles 10 cycles 5 cycles Human Face Recognition 8mm 16mm Maximum sensitivity is centred around 8 to 13 cycles/face width. To recognize with confidence you need to be able to resolve down to 20 cycles/face width

14 1951 USAF Chart Groupings of 6 pairs of bars. Each successive set is half the size of the previous.

15 1951 USAF Chart Groupings of 6 pairs of bars. Each successive set is half the size of the previous. 16mm 8mm

16 Eye charts also provide a simple way of measuring the minimum feature size that can be resolved.

17 20/20 Vision… … or in metric, 6/6 vision Snellen fraction 6 6 Distance at which you should be able to read the line Distance at which you can read the line on the chart Minimum Angle of Resolution

18 Ian Bailey and Jan Lovie The logMAR chart

19 88mm 72mm 58mm 36mm 44mm 6/6 6/12 6/24 6/48 Snellen fraction Letter height Number plate letters 80mm Average eye spacing 65mm 9mm 18mm 6/60 (legally blind)

20 Tests conducted with Pulnix TM6CN 1/2” CCD camera positioned 6m from the target. Images were digitized directly from the camera using a Data Translation 3155 frame grabber C-mount lenses: 4mm 6mm 8.5mm 12.5mm 16mm

21 4mm lens

22 6mm lens

23 8.5mm lens

24 12.5mm lens

25 16mm lens

26 Camera image recorded to video, then played back and digitized. (Look at the USAF chart) Camera image digitized directly. Expect to lose quality when images are recorded to video (cropped images taken with 12.5mm lens)

27 Compression is problematic. Test targets survive compression well, but faces do not. JPEG image quality 0 (14kB)JPEG image quality 4 (24kB) Original PNG image (190kB) JPEG images compressed using Photoshop. Image ‘quality’ can range from 0 - 12

28 JPEG (14kB)JPEG (24kB)Original Faces do not survive compression well

29 What Does Compression Do? Image is divided into 8x8 blocks. Discrete Cosine Transform is applied to each block. The transform coefficients are quantized, many will be rounded to zero. When reconstructed, the amplitude and phase of the spatial frequencies within each 8x8 block will be altered. The 64 basis functions of an 8x8 Discrete Cosine Transform JPEG and MPEG

30 12.5mm lens at 6m No compression ~ 40 pixels

31 12.5mm lens at 6m 18:1 compression

32 12.5mm lens at 6m 18:1 compression

33 12.5mm lens at 6m 31:1 compression

34 12.5mm lens at 6m 31:1 compression 40 pixels across face = 5 DCT blocks Spatial frequencies from 5 cycles/face width upwards are all corrupted This is exactly the range that is most important for face recognition!

35 A Real Surveillance Camera Installation…

36 4.8 m

37

38

39

40 Image quality is defined by many attributes Minimum feature size that can be resolved Noise level Quality of luminance reproduction Quality of colour reproduction.

41 Original laser scanned faces Same shape, varying pigmentation Same pigmentation, varying shape (Russell et al 2007) Luminance and colour cues are at least as important as shape cues People perform about equally well using just shape information or just pigmentation cues.

42 Hue values as greyscale 16 x16 macro-blocks Image compression typically quantizes colour information very heavily…

43 Conclusions Surveillance video, as it is currently used, is almost useless for identification. Face recognition in low resolution images is badly affected by compression artifacts. Image quality standards are needed for surveillance camera installations.

44

45 Conan O’Brien US talk show host Tarja Halonen President of Finland

46

47

48


Download ppt "Video Surveillance: Legally Blind? Peter Kovesi School of Computer Science & Software Engineering The University of Western Australia."

Similar presentations


Ads by Google