Segment dark images by using RAW image data reducing feature noise with LIS
Segment dark images by using RAW image data reducing feature noise with LIS
Instance Segmentation in the Dark
arXiv paper abstract https://arxiv.org/abs/2304.14298
arXiv PDF paper https://arxiv.org/pdf/2304.14298.pdf
Project page https://github.com/Linwei-Chen/LIS
Existing instance segmentation techniques are primarily tailored for high-visibility inputs, but their performance significantly deteriorates in extremely low-light environments.
... The proposed method is motivated by the observation that noise in low-light images introduces high-frequency disturbances to the feature maps of neural networks, thereby significantly degrading performance.
... propose a novel learning method that relies on an adaptive weighted downsampling layer, a smooth-oriented convolutional block, and disturbance suppression learning.
These components effectively reduce feature noise during downsampling and convolution operations, enabling the model to learn disturbance-invariant features.
... discover that high-bit-depth RAW images can better preserve richer scene information in low-light conditions ... can be critical for ... segmentation.
... without any image preprocessing, ... achieve satisfactory performance on instance segmentation in very low light (4% AP higher than state-of-the-art competitors) ...
Please like and share this post if you enjoyed it using the buttons at the bottom!
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
Comments