SDDNet: Infrared small and dim target detection network
Abstract This study focuses on developing deep learning methods for small and dim target detection. We model infrared images as the union of the target region and background region. Based on this model, the target detection problem is considered a two‐class segmentation problem that divides an image...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2023-12-01
|
| Series: | CAAI Transactions on Intelligence Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/cit2.12165 |
| _version_ | 1797383460895064064 |
|---|---|
| author | Ma Long Shu Cong Huang Shanshan Wei Zoujian Wang Xuhao Wei Yanxi |
| author_facet | Ma Long Shu Cong Huang Shanshan Wei Zoujian Wang Xuhao Wei Yanxi |
| author_sort | Ma Long |
| collection | DOAJ |
| description | Abstract This study focuses on developing deep learning methods for small and dim target detection. We model infrared images as the union of the target region and background region. Based on this model, the target detection problem is considered a two‐class segmentation problem that divides an image into the target and background. Therefore, a neural network called SDDNet for single‐frame images is constructed. The network yields target extraction results according to the original images. For multiframe images, a network called IC‐SDDNet, a combination of SDDNet and an interframe correlation network module is constructed. SDDNet and IC‐SDDNet achieve target detection rates close to 1 on typical datasets with very low false positives, thereby performing significantly better than current methods. Both models can be executed end to end, so both are very convenient to use, and their implementation efficiency is very high. Average speeds of 540+/230+ FPS and 170+/60+ FPS are achieved with SDDNet and IC‐SDDNet on a single Tesla V100 graphics processing unit and a single Jetson TX2 embedded module respectively. Additionally, neither network needs to use future information, so both networks can be directly used in real‐time systems. The well‐trained models and codes used in this study are available at https://github.com/LittlePieces/ObjectDetection. |
| first_indexed | 2024-03-08T21:21:19Z |
| format | Article |
| id | doaj.art-6cf0a78d503b4f9ebf6b675290b8c670 |
| institution | Directory Open Access Journal |
| issn | 2468-2322 |
| language | English |
| last_indexed | 2024-03-08T21:21:19Z |
| publishDate | 2023-12-01 |
| publisher | Wiley |
| record_format | Article |
| series | CAAI Transactions on Intelligence Technology |
| spelling | doaj.art-6cf0a78d503b4f9ebf6b675290b8c6702023-12-21T09:45:29ZengWileyCAAI Transactions on Intelligence Technology2468-23222023-12-01841226123610.1049/cit2.12165SDDNet: Infrared small and dim target detection networkMa Long0Shu Cong1Huang Shanshan2Wei Zoujian3Wang Xuhao4Wei Yanxi5School of Computer Science and Engineering Xi'an Technological University Xi'an Shaanxi ChinaSchool of Computer Science and Engineering Xi'an Technological University Xi'an Shaanxi ChinaSchool of Computer Science and Engineering Xi'an Technological University Xi'an Shaanxi ChinaSchool of Computer Science and Engineering Xi'an Technological University Xi'an Shaanxi ChinaSchool of Computer Science and Engineering Xi'an Technological University Xi'an Shaanxi ChinaSchool of Computing University of Kent Canterbury UKAbstract This study focuses on developing deep learning methods for small and dim target detection. We model infrared images as the union of the target region and background region. Based on this model, the target detection problem is considered a two‐class segmentation problem that divides an image into the target and background. Therefore, a neural network called SDDNet for single‐frame images is constructed. The network yields target extraction results according to the original images. For multiframe images, a network called IC‐SDDNet, a combination of SDDNet and an interframe correlation network module is constructed. SDDNet and IC‐SDDNet achieve target detection rates close to 1 on typical datasets with very low false positives, thereby performing significantly better than current methods. Both models can be executed end to end, so both are very convenient to use, and their implementation efficiency is very high. Average speeds of 540+/230+ FPS and 170+/60+ FPS are achieved with SDDNet and IC‐SDDNet on a single Tesla V100 graphics processing unit and a single Jetson TX2 embedded module respectively. Additionally, neither network needs to use future information, so both networks can be directly used in real‐time systems. The well‐trained models and codes used in this study are available at https://github.com/LittlePieces/ObjectDetection.https://doi.org/10.1049/cit2.12165deep learningdetection of moving objects |
| spellingShingle | Ma Long Shu Cong Huang Shanshan Wei Zoujian Wang Xuhao Wei Yanxi SDDNet: Infrared small and dim target detection network CAAI Transactions on Intelligence Technology deep learning detection of moving objects |
| title | SDDNet: Infrared small and dim target detection network |
| title_full | SDDNet: Infrared small and dim target detection network |
| title_fullStr | SDDNet: Infrared small and dim target detection network |
| title_full_unstemmed | SDDNet: Infrared small and dim target detection network |
| title_short | SDDNet: Infrared small and dim target detection network |
| title_sort | sddnet infrared small and dim target detection network |
| topic | deep learning detection of moving objects |
| url | https://doi.org/10.1049/cit2.12165 |
| work_keys_str_mv | AT malong sddnetinfraredsmallanddimtargetdetectionnetwork AT shucong sddnetinfraredsmallanddimtargetdetectionnetwork AT huangshanshan sddnetinfraredsmallanddimtargetdetectionnetwork AT weizoujian sddnetinfraredsmallanddimtargetdetectionnetwork AT wangxuhao sddnetinfraredsmallanddimtargetdetectionnetwork AT weiyanxi sddnetinfraredsmallanddimtargetdetectionnetwork |