Loading...

Efficient Hardware-based Implementation of Object Detection in mmW-Imaging Systems Using AI Algorithms

Gharib, Mohammad Hossein | 2021

206 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54686 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Shabany, Mahdi; Kavehvash, Zahra
  7. Abstract:
  8. Today, due to the increasing activity of terrorist groups, monitoring people in important and busy places such as airports and train stations is very important. One of the technologies that has been developed for this purpose in recent years is 3D imaging technology using millimeter wave. These systems use millimeter waves to image people and identify objects hidden under clothing, which do not have the limitations of conventional imaging techniques such as x-rays and metal detectors. One of the advantages of using these systems is the ability to automatically detect objects in millimeter wave images using deep neural networks such as Segmented, Faster R-CNN and Mask R-CNN, which using these techniques, Increases the speed and accuracy of identifying objects and protecting the privacy of individuals. In this document, various items such as various types of 3D millimeter wave imaging systems, architecture and operation of deep neural networks, instructions for preparing a rich training dataset and data augmentation techniques for preventing overfittnig in neural networks are described and at the end, the results are presented. The results show that the use of data augmentation method proposed in this thesis (data augmentation method by objects extraction) to increase the number and diversity of the training data, significantly increases the accuracy of neural networks. According to the results obtained using this data augmentation method, the accuracy of the Segmented Deep Neural Network is improved from 66.66% to 90.74%, and also Faster R-CNN and Mask R-CNN Deep Neural Networks will have a very good performance in finding the exact objects location in the millimeter wave images
  9. Keywords:
  10. Millimeter Wave ; Artificial Intelligence ; Deep Neural Networks ; Overflowing ; Data Augmentation ; Object Detection

 Digital Object List

 Bookmark

No TOC