Office Document Recognition
By
AI Touch |
Updated hace un mes |
Visual Recognition
Popularity
6.7 / 10
Latency
3,628ms
Service Level
100%
Health Check
N/A
README
Functional Description
- Works with regular document images
- Uploading images returns text recognition results and document area information
| Important featuresClick to apply | Detailed description |
|---|---|
| Applicable Scenarios | Regular document images are supported by default |
| Textual Direction | Supports mixed recognition of horizontal and vertical text by default |
| Printed Handwriting | By default, it supports handwriting and printing judgment and mixed recognition of printed text and handwritten text. |
| Table Analysis | Supports returning table analysis results in the document, such as cell row and column information, etc. |
| Language Type | Default Support Simplified English / Japanese / Korean / French / German / Portuguese / Spanish / Italian / Dutch / Swedish / Finnish / Danish / Norwegian / Hungarian / Vietnamese / Afrikaans / Albanian / Basque / Catalan / Croatian / Czech / Estonian / Icelandic / Irish / Latin/ Chinese / Traditional Chinese / Latvian/Lithuanian/Malay/Polish/Romanian/Slovak/Slovenian/Swahili/Turkish/Welsh/Marathi/Creole/Galician/Esperanto/Filipino/Indonesian/Azerbaijani/Russian/Bulgarian/Macedonian/Ukrainian/Serbian/Belarusian/Greek/Armenian and other 52 common languages. Armenian, 52 common languages and many more mapped languages. |
Request Parameters
The images to be uploaded currently support jpg, png, bmp, pdf, tiff, single frame gif and most other formats. The file size should not exceed 10M, and the width and height of the image should be between 20 and 10000 (pixels).
| parameter name | data type | description |
|---|---|---|
| imageUrl | String | Valid URL or base64 string starting with “http(s)😕/” |
| imageFile | File |
Request Body
Content-Type: application/json
The JSON structure is described below:
Response Format
| Field | Type | Description |
|---|---|---|
| code | string | Response code, “200” for success |
| status | string | Response status, “SUCCESS” for success |
| service | string | Service name, “office_documents_ocr” here |
| date | string | Response date and time |
| result | object | Recognition result object |
Recognition Result Object (result)
| Field | Type | Description |
|---|---|---|
| angle | number | Angle of the whole image |
| height | number | Image height |
| lines | array | Recognition results by text lines |
| areas | array | Recognized areas array |
| tables | array | Recognized tables array |
| width | number | Image width |
Text Line Object (lines)
| Field | Type | Description |
|---|---|---|
| angle | number | This field represents the rotation angle of an input image in clockwise direction from the upright position, defined as 0 degrees, which corresponds to the orientation of images when humans read text. 0: ▲ (upright position) 90: ▶ (rotated 90 degrees clockwise) 180: ▼ (rotated 180 degrees clockwise) 270: ◀ (rotated 270 degrees clockwise) |
| area_index | number | Area index, same if in same paragraph, output when apply_layout=1 |
| area_type | string | Area type, output when apply_layout=1 |
| direction | number | Text reading direction -1: other 0: single character 1: horizontal 2: vertical |
| handwritten | number | Whether the text is handwritten or not. -1: Unknown 0: Text is not handwritten, usually printed 1: Text is handwritten, usually with distinctive writing characteristics |
| position | array | The coordinates of the identified value in the transitive map are an array of length 8 [0,1,2,3,4,5,6,7]. (0, 1) coordinates of the upper left corner (2, 3) coordinates of the upper right corner (4, 5) Coordinate of the lower right corner. (6, 7) Bottom-left corner coordinates |
| score | number | Confidence score |
| text | string | Recognized text |
| type | string | Text type, used to represent the form of the text. Under the current version, the text type includes: text (text) stamp (stamp) formula (formula) |
Area Object (areas)
| Field | Type | Description |
|---|---|---|
| index | number | Index of area |
| position | array | The coordinates of the identified value in the transitive map are an array of length 8 [0,1,2,3,4,5,6,7]. (0, 1) coordinates of the upper left corner (2, 3) coordinates of the upper right corner (4, 5) Coordinate of the lower right corner. (6, 7) Bottom-left corner coordinates |
| score | number | Confidence score |
| type | string | Area types: paragraph list edge image stamp formula watermark table borderless_table |
| text | string | Text contained in the area |
Table Object (tables)
| Field | Type | Description |
|---|---|---|
| area_index | number | Area index, same if in same paragraph, output when apply_layout=1 |
| area_type | string | Area type, output when apply_layout=1 |
| lines | array | Text lines, for non-table areas |
| position | array | The coordinates of the identified value in the transitive map are an array of length 8 [0,1,2,3,4,5,6,7]. (0, 1) coordinates of the upper left corner (2, 3) coordinates of the upper right corner (4, 5) Coordinate of the lower right corner. (6, 7) Bottom-left corner coordinates |
| table_cells | array | Table cells, for lined/lineless tables |
| table_cols | number | Number of columns |
| table_rows | number | Number of rows |
| type | string | Table types: plain - text area table_with_line - lined table table_without_line - lineless table |
| width_of_cols | array | Width of each column when table is upright |
| height_of_rows | array | Height of each row when table is upright |
Followers: 0
API Creator:
AI Touch
dominonet-lTpEE6zONeS
Log In to Rate API
Rating: 5 - Votes: 1