Se-DIFT predicts feature appearances across RGB and thermal modalities via an encoder-decoder plus global feature vector, cutting L1 error over 7% versus U-Net and enabling intermodal matching of SIFT, SURF, and ORB.
Visual Multimodal Odometry: Robust Visual Odometry in Harsh Environments
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Semantic Deep Intermodal Feature Transfer: Transferring Feature Descriptors Between Imaging Modalities
Se-DIFT predicts feature appearances across RGB and thermal modalities via an encoder-decoder plus global feature vector, cutting L1 error over 7% versus U-Net and enabling intermodal matching of SIFT, SURF, and ORB.