search

khurramHashmi / FeatEnHancer

[ICCV 2023] FeatEnHancer: Enhancing Hierarchical Features for Object Detection and Beyond Under Low-Light Vision
https://khurramhashmi.github.io/FeatEnHancer/
Other
15 stars 0 forks source link

"self.scale_aware_aggregation = ScaleAwareFeatureAggregation(channels=32).to('cuda')" #3

Closed Felix-fz closed 1 month ago

Felix-fz commented 2 months ago

Hi, Thanks for your insteresting work! But when I run the "FeatEnHancer", the "channels" of"self.scale_aware_aggregation = ScaleAwareFeatureAggregation(channels=32).to('cuda')" is wrong? Because the "out_channels=32". And the result of "self.quarter_conv" is [B, int_out_channels, H/4-1,W/4-1]? image Best wishes!

khurramHashmi commented 1 month ago

Hi,

Thank you for your interest in our work, and apologies for the late response.

The self.quarter_conv is used to downscale the input image, and its output is not utilized in ScaleAwareFeatureAggregation (SAFA). However, you are correct that there is a mismatch with channels=32 in SAFA, which should be channels=24. This has now been resolved. Thank you for bringing this up.