search

zhiyuanyou / SAFECount

[WACV 2023] Few-shot Object Counting with Similarity-Aware Feature Enhancement
Other
124 stars 22 forks source link

question on the use of few shot images #35

Closed limshanshan closed 5 months ago

limshanshan commented 6 months ago

Hi, may I know for the FSC147 to CARPK experiment, after I pretrain the model, must I finetune it or can I go straight to eval? If I finetune it, won't the unseen class be exposed to the model? Also, is there support images used during eval or finetuning?

zhiyuanyou commented 5 months ago
  1. Both pre-trained model and fine-tuned model are tested.
  2. Yes. The unseen class will be exposed to the model during the fine-tuning stage.
  3. Yes. There are support images used during eval or fine-tuning.
limshanshan commented 5 months ago

ok thank u :)

limshanshan commented 5 months ago

Hi, for FSC147 to CARPK experiment, do I have to manually exclude the car images during training using FSC147?

zhiyuanyou commented 5 months ago

Yes.

limshanshan commented 5 months ago

ok, thank u!