A STUDY ON APPLICATION OF MACHINE LEARNING AND COMPUTER VISION FOR RETAIL PROJECTS
DOI:
https://doi.org/10.22159/ajpcr.2017.v10s1.20522Keywords:
Computer vision, Neural network, Convolutional neural network, Deep learning, Inception architectureAbstract
The motivation for developing computer vision is the human vision system which is the richest sense that we have. To us, vision seems an easy task of
just seeing objects in daily life and identifying them, but in reality, our eyes along with the brain are processing information of around 50 images every
second with millions of pixels in each image. Most of these images obtained are currently just looked at by people. The challenging task is to process
images from all these cameras and allow automation of tasks never before considered. Neural networks help us in making cameras intelligent enough
to understand the images it captures. Convolutional neural networks (CNN) are trained to give image classification results of good accuracy, with the
challenge to improve utilization of computing resources. Google Net is in its essence a deep CNN that uses inception architecture to attain leading
edge results for classification and detection problems. In this paper, a study was made on applications of computer vision techniques in retail and
customer strategic projects. Further, it was analyzed that if cameras trained with CNN can work well enough to be deployed in retail market scenarios
to automate sales and stock supervision.
Downloads
References
Computer Vision. Available from: https://www.en.wikipedia.org/wiki/from: http://www.cs231n.
stanford.edu/.Computer_vision.2. Convolutional Neural Networks for Visual Recognition. Available
from: http://www.cs231n.stanford.edu/.
Convolutional Neural Networks (LeNet). Available online: http:// deeplearning.net/tutorial/lenet.html.
Domingos P. A Few Useful Things to Know about Machine Learning. Vol. 55. University of Washington: Communications of the ACM; 2012. p. 78-87.
Hinton G, Krizhevsky A, Srivastava N. Improving neural networks by preventing co-adaption of feature detectors, arXiv:1207.0580v1 [cs. NE] 3 July; 2012.
Szegedy C, Liu W, Jia Y, Sermanet P, Reed S. Going deeper with convolutions. Computer Vision and Pattern Recognition, arXiv:1409.4842v1 [cs.CV] 17 September; 2014.
Szegedy C, Vanhouchke V, Ioffe S, Shlens J, Wojna Z. Rethinking the Inception Architecture for Computer Vision, arXiv:1512.00567v3 [cs. CV] 11 December; 2015.
Lin M, Chen Q, Yan S. Network in Network. arXiv:1312.4400v3 [cs. NE] 4 March; 2014.
Srivastava N, Hinton G, Krizhevsky A. Dropout: A simple way to prevent neural networks from over fitting. J Mach Learn Res 2014;15(1):1929-58.
Nielsen M. Neural Networks and Deep Learning. Available from:http:// www.neuralnetworksan
ddeeplearning.com/index.html.
Convolutional Neural Networks for Visual Recognition. Available form: http://www.cs231
n.stanford.edu/.
Building Powerful Image Classification Models using Very Little Data. Available from: https://www.blog.keras.io/building-powerful-imageclassification- models-using-very-little-data.html.
Deep Learning Framework: Caffe. Available from: http://www.caffe. berkeleyvision.org/.
Inceptionism: Going Deeper into Neural Networks. Available from: https://www.research.
googleblog.com/2015/06/inceptionism-goingdeeper- into-neural.html.
TensorFlow Playground. Available from: http://www.playground. tensorflow.org.
Krizhevsky A, Sutskever I, Hinton GE. Image Net Classification with Deep Convolutional Neural Networks. Advances in Neural Information Proceeding Systems (NIPS; 2012).
Published
How to Cite
Issue
Section
The publication is licensed under CC By and is open access. Copyright is with author and allowed to retain publishing rights without restrictions.
