Please use this identifier to cite or link to this item:
Title: Efficient Object Detection using Transfer Learning
Authors: Singh, Harshdeep
Supervisor: Sharma, R. K.
Keywords: Object Detection;Transfer Learning;VGG-16;ImageNet
Issue Date: 9-Aug-2019
Abstract: Object detection in automated surveillance images or video is an extremely monotonous process for monitoring for crowded scenes and a variety of sensible objects can be restricted in surveillance images or video. An appropriate machine learning technique can help to train the object detection system in identifying random activities during surveillance. To this end, we present Efficient Object Detection using Transfer Learning that can be used as a tool for object detection in surveillance images or videos using the concept of artificial intelligence. The main intention of the proposed object detection system is to improve the detection time and accuracy by using the concept of Convolutional Neural Network (CNN) as artificial intelligence technique. In this paper we present CNN based VGG-16 model for object detection, which is the combination of multiple layer of hidden unit with the optimized feature by using transfer learning. Here CNN is used for classifying the random activity into objects from the surveillance images or videos based on the transfer learning which is used for the selection of optimal feature sets. Further, Self adaptive transfer learning is adopted to efficiently solve optimization problems in the continuous search domain to select the best possible feature to segregate the pattern of object. The main contribution of this research is validation of proposed system for the large scale data and we introduce a new large-scale dataset of 50 class images. Dataset consists of total 5000 long and untrimmed real-world surveillance indoor images. The experimental results of the proposed system show that our designed for object detect achieve significant improvement on detection system performance as compared to the state-of-the-art approaches. In this paper, to validate the proposed model we provide the comparison of existing results of several recent deep learning baselines on object detection. The real-time object detection in surveillance images or videos sequences using transfer learning based on CNN with feature extraction technique is implemented using anaconda3 python Software.
Appears in Collections:Masters Theses@CSED

Files in This Item:
File Description SizeFormat 
Thesis 801732019.pdf1.7 MBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.