Presentation is loading. Please wait.

Presentation is loading. Please wait.

Pace DPS Semantic Geometric Features: A Preliminary Investigation of Automobile Identification Carl E. Abrams Sung-Hyuk Cha, Michael Gargano, and Charles.

Similar presentations


Presentation on theme: "Pace DPS Semantic Geometric Features: A Preliminary Investigation of Automobile Identification Carl E. Abrams Sung-Hyuk Cha, Michael Gargano, and Charles."— Presentation transcript:

1 Pace DPS Semantic Geometric Features: A Preliminary Investigation of Automobile Identification Carl E. Abrams Sung-Hyuk Cha, Michael Gargano, and Charles Tappert

2 Pace DPS Agenda Overview of the Problem The Experiments Results Going Forward

3 Pace DPS Overview Object recognition remains a hard problem The human mind uses shapes to recognize objects Can semantic features defined by their shapes be more effective in the recognition and identification of objects?

4 Pace DPS The Experiments 10 test images of cars Directly form the manufactures websites Images were restricted to side views of the cars taken from 90 degrees All 2005 models Feature vectors calculated/measured from the images

5 Pace DPS The Vehicles

6 Pace DPS Experiments used Euclidean Distance as the Measure the xi and ti are measurements from two different vehicles

7 Pace DPS Experiments used Euclidean Distance as the Measure (x1,y1) (x2,y2) c a b c = (a 2 +b 2 ) 1/2 c = ((x1-x2) 2 +(y1-y2) 2 ) 1/2

8 Pace DPS Manufacturers Specifications First Experiment

9 Pace DPS Boundary Description using Rays Second Experiment

10 Pace DPS Semantic Features Third Experiment

11 Pace DPS Challenge: Determine the qualitative ability of the feature vectors to separate the vehicles Within each experiment compute the distance of each vehicle from all the others Evenly divide the measures into 5 bins Observe the distribution of the measures

12 Pace DPS The Results

13 Pace DPS Distance Matrix – Semantic Features 0156.02153.17154.0027.66155.962.82151.9026.07156.23 156.0207.217.07161.7216.09156.3613.45154.014.12 153.177.2104.24159.7811.44153.609.43150.097.55 154.007.074.240160.899.43154.376.55151.065.00 27.65161.73159.77160.890164/3527.58159.8451.08162.50 155.9616.0911.449.43164.350156.365.83151.6313.34 2.82156.36153.60154.3627.58156.360152.2428.00156.53 151.9013,459.436.55159.845.83152.240148.3310.48 26.07154.01150.09151.0651.07151.6328.00148.330154.01 156.234.127.555.00162.5013.34156.5310.48154.010 Honda Civic Honda Accord Mazda 3Mazda 6Porsche Carerra Toyota Camry Toyota Celica Toyota Corolla Toyota Echo VW Passat Honda Civic Honda Accord Mazda 3 Mazda 6 Porsche Carerra Toyota Camry Toyota Celica Toyota Corolla Toyota Echo VW Passat

14 Pace DPS Going Forward Extend techniques to encompass semantic shapes within an object (shape contexts) Compare the extended semantic methods to existing methods in multiple domains

15 Pace DPS Going Forward Shape Contexts

16 Pace DPS References [1]R. D. Acqua and R. Job, "Is global shape sufficient for automatic object identification?" Congitive Science, vol. 8, pp. 801- 821, 2001. [2]A. K. Jain, A. Ross, and S. Pankanti, "A Prototype Hand Geomtery-based Verification System," presented at Proceedings of 2nd International conference on Audio and Video-based Biometric Person Authentication, Wahington D.C., 1999. [3]H. Schneiderman and T. Kanade, "A Statistical Model for 3D Object Detection Applied to Faces and Cars," presented at IEEE Conference on Computer Vision and Pattern Recognition, IEEE, 2000 [4] S. Belongie,J Malik, J Puzicha, “Matching Shapes”,presented at the International Conference on Computer Vision (ICCV 01) Vol 1, Jan 2001


Download ppt "Pace DPS Semantic Geometric Features: A Preliminary Investigation of Automobile Identification Carl E. Abrams Sung-Hyuk Cha, Michael Gargano, and Charles."

Similar presentations


Ads by Google