faster-r-cnn-neutron-images
Science Score: 44.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (5.7%) to scientific vocabulary
Repository
Basic Info
- Host: GitHub
- Owner: anedicksh
- Language: Jupyter Notebook
- Default Branch: main
- Size: 3.91 MB
Statistics
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
Faster R-CNN algorithm for object detection in neutron images
Nowadays, inspection of baggage for threat objects (such as explosives) using images is of interest as it can contribute towards more effective and advanced security screen- ing systems. As neutron imaging presents certain advantages over traditional X-Ray imaging, Dynaxion is a company located in Eindhoven that designed a screening sys- tem based on neutron generation. Moreover, the checking of baggage is partially based on manual detection. While this can be efficient, it is also costly and leaves window open for human error. Therefore, this study aims at optimizing a Faster R-CNN that can accurately detect and classify illegal objects contained in neutron images that were generated from Dynaxion. For this, six Faster R-CNN models were trained based on different techniques of feature extraction (ResNet50 and MobileNet-v3), optimiz- ers (Adam or SGD), learning rates, batch sizes, number of epochs and weight decays. In conclusion, the best detection was achieved by the combination of the SGD opti- mizer and the ResNet50 feature extractor, achieving an overall F1 score of 0.85. Sample images on a test dataset demonstrate the model’s accurate object detection capabilities.
Owner
- Login: anedicksh
- Kind: user
- Repositories: 1
- Profile: https://github.com/anedicksh
Citation (CITATION.cff)
cff-version: 1.2.0 message: "If you use this software, please cite it as below." authors: - family-names: "Dicks Herrán" given-names: "Ane" orcid: "https://orcid.org/0000-0000-0000-0000" title: "Faster R-CNN in neutron images" version: 2.0.4 doi: 10.5281/zenodo.1234 date-released: 2017-12-18 url: "https://github.com/anedicksh/Faster-R-CNN-neutron-images"