This folder contains the evaluation code for GPG2A Ground-to-Aerial (G2A) image synthesis model.

Prerequisite

pytorch
torchvision
scikit-image
opencv
lpips
PIL
tqdm

Pre-trained SAFA model

Please download the pretrained SAFA model here and extract to the root directory.

Generate aerial images

Please follow the training and testing tutorial to generate aerial images for evaluation with correct naming and directory.

$Sim_c$ and $Sim_s$

To calculate $Sim_c$ and $Sim_s$, please run the command below,

python eval_SAFA.py --model_path SAFA_PRETREAINED_MODEL_PATH --experiment_name YOUR_PREFERRED_EXPERIMENT_NAME --image_path GENERATED_IMAGE_PATH --gt_path GROUND_TRUTH_IMAGE_PATH

For example,

python eval_SAFA.py --model_path ../GPG2A_eval_SAFA_model/ --experiment_name GPG2A_experiment --image_path ../log_imgs_test/ --gt_path ../Data/VIGOR/

FID-SAFA

To calculate FID-SAFA, please run the command below,

python eval_FID.py --model_path SAFA_PRETREAINED_MODEL_PATH --experiment_name YOUR_PREFERRED_EXPERIMENT_NAME --image_path GENERATED_IMAGE_PATH --gt_path GROUND_TRUTH_IMAGE_PATH

For example,

python eval_FID.py --model_path ../GPG2A_eval_SAFA_model/ --experiment_name GPG2A_experiment --image_path ../log_imgs_test/ --gt_path ../Data/VIGOR/

LPIPS, SSIM, PSNR

To calculate LPIPS, SSIM, PSNR, please run the command below,

python test_lpips.py --experiment_name YOUR_PREFERRED_EXPERIMENT_NAME --image_path GENERATED_IMAGE_PATH --gt_path GROUND_TRUTH_IMAGE_PATH

For example,

python test_lpips.py --experiment_name GPG2A_experiment --image_path ../log_imgs_test/ --gt_path ../Data/VIGOR/