Radio Galaxy Detection - Computer Vision Algorithms
CSIRO RDS Repository
View Archive InfoField | Value | |
Title |
Radio Galaxy Detection - Computer Vision Algorithms
|
|
Creator |
Nikhel Gupta
|
|
Subject |
Machine learning not elsewhere classified
|
|
Description |
This collection of machine-learning algorithms for detecting radio sources and their infrared host galaxies. We introduce Gal-DETR, Gal-Deformable DETR and Gal-DINO multimodal models for object detection. These models are built upon the DETR (Carion et al., 2020), Deformable DETR (Zhu et al., 2021), and DINO (Zhang et al., 2022) algorithms, which are optimal methods for predicting bounding box instances and categories of required objects. We extended their capabilities by incorporating keypoint detection. In addition to the bounding boxes employed to detect extended radio galaxies, the integration of keypoint detection techniques offers a complementary approach for identifying infrared hosts. For more details, see the RadioGalaxyNET paper in the PASA journal and the NeurIPS 2023 conference workshop. |
|
Publisher |
CSIRO
|
|
Contributor |
Zeeshan Hayder
Ray Norris Minh Huynh Lars Petersson |
|
Date |
2023-12-01
|
|
Type |
—
|
|
Format |
—
|
|
Identifier |
csiro:61069
|
|
Language |
—
|
|
Coverage |
—
|
|
Rights |
—
|
|