search

mit-quest / necstlab-damage-segmentation

MIT License
5 stars 6 forks source link

To `infer_segmentation.py`, add input parameter to command that toggles data output type(s) #30

Closed rak5216 closed 4 years ago

rak5216 commented 4 years ago

We want to choose what data output we get from infer. Currently we only get a 8-bit RGB with damage overlayed on the raw data, which is at least a 3X larger file vs. inputs. We would like an input parameter that acts as a flag to establish whether we want (1) mask only, (2) mask+raw , or (3) both stacks.

rak5216 commented 4 years ago 
pipenv run python3 infer_segmentation.py --gcp-bucket gs://necstlab-sandbox --model-id segmentation-model-small_20190924T191717Z --stack-id THIN_REF_S2_P1_L3_2496_1563_2159 --image-ids 8bit_AS4_CNT_S1_P1_L2_2560_1750_2160_0092_neg_class_0_crop0.tif,8bit_AS4_CNT_S1_P1_L2_2560_1750_2160_0092_neg_class_0_crop0.tif,8bit_AS4_CNT_S1_P1_L2_2560_1750_2160_0092_neg_class_0_crop0.tif --output overlay
rak5216 commented 4 years ago

or

pipenv run python3 infer_segmentation.py --gcp-bucket gs://necstlab-sandbox --model-id segmentation-model-small_20190924T191717Z --stack-id THIN_REF_S2_P1_L3_2496_1563_2159 --image-ids 8bit_AS4_CNT_S1_P1_L2_2560_1750_2160_0092_neg_class_0_crop0.tif,8bit_AS4_CNT_S1_P1_L2_2560_1750_2160_0092_neg_class_0_crop0.tif,8bit_AS4_CNT_S1_P1_L2_2560_1750_2160_0092_neg_class_0_crop0.tif --output labels