出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：Evolution of visual object recognition architectures based on Convolutional Neural Networks &Convolutional Deep Belief Networks paradigms has revolutionized artificial Vision Science.These architectures extract & learn the real world hierarchical visual features utilizingsupervised & unsupervised learning approaches respectively. Both the approaches yet cannotscale up realistically to provide recognition for a very large number of objects as high as 10K.We propose a two level hierarchical deep learning architecture inspired by divide & conquerprinciple that decomposes the large scale recognition architecture into root & leaf level modelarchitectures. Each of the root & leaf level models is trained exclusively to provide superiorresults than possible by any 1-level deep learning architecture prevalent today. The proposedarchitecture classifies objects in two steps. In the first step the root level model classifies theobject in a high level category. In the second step, the leaf level recognition model for therecognized high level category is selected among all the leaf models. This leaf level model ispresented with the same input object image which classifies it in a specific category. Also wepropose a blend of leaf level models trained with either supervised or unsupervised learningapproaches. Unsupervised learning is suitable whenever labelled data is scarce for the specificleaf level models. Currently the training of leaf level models is in progress; where we havetrained 25 out of the total 47 leaf level models as of now. We have trained the leaf models withthe best case top-5 error rate of 3.2% on the validation data set for the particular leaf models.Also we demonstrate that the validation error of the leaf level models saturates towards theabove mentioned accuracy as the number of epochs are increased to more than sixty. The top-5error rate for the entire two-level architecture needs to be computed in conjunction with theerror rates of root & all the leaf models. The realization of this two level visual recognitionarchitecture will greatly enhance the accuracy of the large scale object recognition scenariosdemanded by the use cases as diverse as drone vision, augmented reality, retail, image search& retrieval, robotic navigation, targeted advertisements etc.
关键词：Convolutional Neural Network [CNN]; Convolutional Deep Belief Network [CDBN];Supervised & Unsupervised training
