|
| 1 | +#### Description about SROIE datalist: |
| 2 | +The original [SROIE](https://rrc.cvc.uab.es/?ch=13&com=introduction) dataset contains many noise. Here, we release a clean version of SROIE datalist for fair comparison with our TRIE model. |
| 3 | +The SROIE datalist includes OCR and IE annotations. |
| 4 | + |
| 5 | +- bboxes: four points annotation of text boxes(top-left, top-right, bottom-right, bottom-left) |
| 6 | + |
| 7 | +- texts: text content annotation |
| 8 | + |
| 9 | +- cares: whether each box is noise or not(0 means noise and will be ignored during training or testing) |
| 10 | + |
| 11 | +- labels: labels of each box (0 for 'other', 1 for 'company', 2 for 'total', 3 for 'total', 4 for 'date', 5 for 'address') |
| 12 | + |
| 13 | +- bbox_bieo_labels: labels of each character in a box (with same meanings as labels) |
| 14 | + |
| 15 | + |
| 16 | + |
| 17 | +#### Datalists: |
| 18 | + |
| 19 | + davar_train_datalist_w_bieo.json: 626 training samples |
| 20 | + |
| 21 | + davar_test_datalist_w_bieo.json: 347 testing samples |
| 22 | + |
| 23 | + classes_config is as follows: |
| 24 | +``` |
| 25 | +{ |
| 26 | + "classes": |
| 27 | + [ |
| 28 | + "others", |
| 29 | + "company", |
| 30 | + "total", |
| 31 | + "date", |
| 32 | + "address" |
| 33 | + ] |
| 34 | +} |
| 35 | +``` |
| 36 | + |
| 37 | + |
| 38 | + |
| 39 | +#### Example: |
| 40 | + |
| 41 | + { |
| 42 | + "Images/train/X51005757349.jpg": |
| 43 | + { |
| 44 | + "height": 1373, |
| 45 | + "width": 703, |
| 46 | + "content_ann": |
| 47 | + { |
| 48 | + "bboxes": [[227, 159, 457, 159, 457, 185, 227, 185], |
| 49 | + [242, 193, 426, 193, 426, 223, 242, 223], |
| 50 | + [185, 229, 482, 229, 482, 262, 185, 262], |
| 51 | + [198, 267, 481, 267, 481, 299, 198, 299], |
| 52 | + [170, 305, 499, 305, 499, 337, 170, 337], |
| 53 | + [226, 343, 454, 343, 454, 372, 226, 372], |
| 54 | + ... |
| 55 | + ], |
| 56 | + "cares": [1, 1, 1, 1, 1, 1, ...], |
| 57 | + "labels": [[1], |
| 58 | + [0], |
| 59 | + [4], |
| 60 | + [4], |
| 61 | + [4], |
| 62 | + [0], |
| 63 | + ... |
| 64 | + ], |
| 65 | + "bbox_bieo_labels":[[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], |
| 66 | + [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], |
| 67 | + [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4], |
| 68 | + [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4], |
| 69 | + [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4], |
| 70 | + [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], |
| 71 | + ... |
| 72 | + ] |
| 73 | + "texts": [ |
| 74 | + "GOLDEN KEY MAKER", |
| 75 | + "(000760274-K)", |
| 76 | + "NO 5, JALAN KENARI 2,", |
| 77 | + "BANDAR PUCHONG JAYA,", |
| 78 | + "47100 PUCHONG, SELANGOR", |
| 79 | + "TEL: 03-58919941", |
| 80 | + ... |
| 81 | + ] |
| 82 | + } |
| 83 | + }, |
| 84 | + ... |
| 85 | + } |
| 86 | + |
| 87 | + |
| 88 | +##### Download Link: |
| 89 | + |
| 90 | +You can download the datalists via this [link](https://one.hikvision.com/#/link/O0DXYBPhlqpGQI7nmRFA). (Access Code: XYAZ) |
| 91 | + |
0 commit comments