Shopping carts have traditionally been used as a tool provided to the customers in retail stores to carry items from the shelf to checkout stations. These days shopping carts can also be used as a security checkpoint to prevent store losses. All the items collected in a shopping cart are supposed to be unloaded at the checkout station to be scanned and included in the bill. Any items left in the cart intentionally or by accident will not be charged and therefore cause a loss to the store. We propose a system that automatically detects shopping carts and verify their emptiness at the checkout station. We use motion segmentation, line detection, and template matching methods for the cart detection and emptiness verification. An inter-frame edge difference, cart’s path accumulator, and a finite state model are introduced for accurate cart detection. All detected carts are compared with empty cart models and the dissimilarity scores are calculated to verify the emptiness. The proposed system was evaluated on a long video clip (~12 hours) and showed promising results both in cart detection and emptiness verification.
[1]
P. J. Burt,et al.
Change Detection and Tracking Using Pyramid Transform Techniques
,
1985,
Other Conferences.
[2]
Rita Cucchiara,et al.
Detecting Moving Objects, Ghosts, and Shadows in Video Streams
,
2003,
IEEE Trans. Pattern Anal. Mach. Intell..
[3]
Roberto Brunelli,et al.
Advanced
,
1980
.
[4]
Richard O. Duda,et al.
Pattern classification and scene analysis
,
1974,
A Wiley-Interscience publication.
[5]
Matthijs C. Dorst.
Distinctive Image Features from Scale-Invariant Keypoints
,
2011
.
[6]
William B. Thompson,et al.
Detecting moving objects
,
1989,
International Journal of Computer Vision.
[7]
Roberto Brunelli,et al.
Template Matching Techniques in Computer Vision: Theory and Practice
,
2009
.
[8]
Jan A Snyman,et al.
Practical Mathematical Optimization: An Introduction to Basic Optimization Theory and Classical and New Gradient-Based Algorithms
,
2005
.