My F1Score is 0 and pixelRUROC also 0? Why?

[lofangmeng]

My custom datasets was prepared same like the MvTec Format

but the result come out with no ground-truth and predict

Is that any wrong on my datasets or what?

[samet-akcay]

@lofangmeng, could you please share your config file for reference? Thanks?

[lofangmeng]

dataset:
name: mvtec #options: [mvtec, btech, folder]
format: mvtec
path: ./datasets/MVTec
task: segmentation
category: makoutie
image_size: 224
train_batch_size: 32
test_batch_size: 32
num_workers: 0
transform_config:
train: null
val: null
create_validation_set: false
tiling:
apply: false
tile_size: null
stride: null
remove_border_count: 0
use_random_tiling: False
random_tile_count: 16

model:
name: patchcore
backbone: wide_resnet50_2
pre_trained: true
layers:
- layer2
- layer3
coreset_sampling_ratio: 0.1
num_neighbors: 9
normalization_method: min_max # options: [null, min_max, cdf]

[lofangmeng]

i was using the MVTec format

[lofangmeng]

are the ground trouh should be white color? cannot be other color?

[samet-akcay]

are the ground trouh should be white color? cannot be other color?

the datamodule would expect a binary mask of 1's and 0's, where 1 (white) would be the anomalous region. You could refer to an MVTec ground-truth image.

[ashwinvaidya17]

@lofangmeng were you able to get the results? If possible can you share an example of the ground truth masks?