---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: mask
    dtype: image
  - name: label
    dtype: int64
  splits:
  - name: candle.train
    num_bytes: 106451773.0
    num_examples: 900
  - name: candle.test
    num_bytes: 23359449.0
    num_examples: 200
  - name: capsules.train
    num_bytes: 133021141.0
    num_examples: 542
  - name: capsules.test
    num_bytes: 39865980.0
    num_examples: 160
  - name: cashew.train
    num_bytes: 135528457.0
    num_examples: 450
  - name: cashew.test
    num_bytes: 48713873.0
    num_examples: 150
  - name: chewinggum.train
    num_bytes: 63491934.0
    num_examples: 453
  - name: chewinggum.test
    num_bytes: 21472874.0
    num_examples: 150
  - name: fryum.train
    num_bytes: 63780392.0
    num_examples: 450
  - name: fryum.test
    num_bytes: 21646212.0
    num_examples: 150
  - name: macaroni1.train
    num_bytes: 105415318.0
    num_examples: 900
  - name: macaroni1.test
    num_bytes: 24090768.0
    num_examples: 200
  - name: macaroni2.train
    num_bytes: 100349144.0
    num_examples: 900
  - name: macaroni2.test
    num_bytes: 22470288.0
    num_examples: 200
  - name: pcb1.train
    num_bytes: 244978923.0
    num_examples: 904
  - name: pcb1.test
    num_bytes: 53521326.0
    num_examples: 200
  - name: pcb2.train
    num_bytes: 224276308.0
    num_examples: 901
  - name: pcb2.test
    num_bytes: 51075179.0
    num_examples: 200
  - name: pcb3.train
    num_bytes: 127418394.0
    num_examples: 905
  - name: pcb3.test
    num_bytes: 28467534.0
    num_examples: 201
  - name: pcb4.train
    num_bytes: 192400641.0
    num_examples: 904
  - name: pcb4.test
    num_bytes: 44329307.0
    num_examples: 201
  - name: pipe_fryum.train
    num_bytes: 42230565.0
    num_examples: 450
  - name: pipe_fryum.test
    num_bytes: 14593580.0
    num_examples: 150
  download_size: 1917906073
  dataset_size: 1932949360.0
configs:
- config_name: default
  data_files:
  - split: candle.train
    path: data/candle.train-*
  - split: candle.test
    path: data/candle.test-*
  - split: capsules.train
    path: data/capsules.train-*
  - split: capsules.test
    path: data/capsules.test-*
  - split: cashew.train
    path: data/cashew.train-*
  - split: cashew.test
    path: data/cashew.test-*
  - split: chewinggum.train
    path: data/chewinggum.train-*
  - split: chewinggum.test
    path: data/chewinggum.test-*
  - split: fryum.train
    path: data/fryum.train-*
  - split: fryum.test
    path: data/fryum.test-*
  - split: macaroni1.train
    path: data/macaroni1.train-*
  - split: macaroni1.test
    path: data/macaroni1.test-*
  - split: macaroni2.train
    path: data/macaroni2.train-*
  - split: macaroni2.test
    path: data/macaroni2.test-*
  - split: pcb1.train
    path: data/pcb1.train-*
  - split: pcb1.test
    path: data/pcb1.test-*
  - split: pcb2.train
    path: data/pcb2.train-*
  - split: pcb2.test
    path: data/pcb2.test-*
  - split: pcb3.train
    path: data/pcb3.train-*
  - split: pcb3.test
    path: data/pcb3.test-*
  - split: pcb4.train
    path: data/pcb4.train-*
  - split: pcb4.test
    path: data/pcb4.test-*
  - split: pipe_fryum.train
    path: data/pipe_fryum.train-*
  - split: pipe_fryum.test
    path: data/pipe_fryum.test-*
---


Original dataset:

```
@inproceedings{zou2022spot,
  title={Spot-the-difference self-supervised pre-training for anomaly detection and segmentation},
  author={Zou, Yang and Jeong, Jongheon and Pemula, Latha and Zhang, Dongqing and Dabeer, Onkar},
  booktitle={European Conference on Computer Vision},
  pages={392--408},
  year={2022},
  organization={Springer}
}
```