You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: articles/cognitive-services/form-recognizer/whats-new.md
+65-4Lines changed: 65 additions & 4 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -8,19 +8,80 @@ manager: nitinme
8
8
ms.service: cognitive-services
9
9
ms.subservice: forms-recognizer
10
10
ms.topic: conceptual
11
-
ms.date: 12/12/2019
11
+
ms.date: 03/20/2020
12
12
ms.author: pafarley
13
13
14
14
---
15
15
16
16
# What's new in Form Recognizer?
17
17
18
-
This article highlights the major changes that come with new versions of the Form Recognizer API.
18
+
The Form Recognizer service is updated on an ongoing basis. Use this article to stay up to date with feature enhancements, fixes, and documentation updates.
19
19
20
20
> [!NOTE]
21
-
> The quickstarts and guides in this doc set always use the latest version of the API, unless they specify otherwise.
21
+
> The quickstarts and guides for Form Recognizer always use the latest version of the API, unless specified.
22
22
23
-
## Form Recognizer 2.0 (preview)
23
+
## March 2020
24
+
25
+
### Extraction Enhancements and bug fixes
26
+
27
+
This release includes extraction enhancements and accuracy improvements, specifically, the capability to label and extract multiple key/value pairs in the same line of text. In addition, we've fixed several bugs, including a fix that allows users to submit a file URL as the source to the [`analyzeForm`](https://westus2.dev.cognitive.microsoft.com/docs/services/form-recognizer-api-v2-preview/operations/AnalyzeWithCustomForm) and [`analyzeLayout`](https://westus2.dev.cognitive.microsoft.com/docs/services/form-recognizer-api-v2-preview/operations/AnalyzeLayoutAsync) APIs.
28
+
29
+
30
+
**Enhancements**
31
+
32
+
* Added additional logs and telemetry for storage manager, time-to-results, and document complexity
33
+
* Improved model performance
34
+
* Integrated latest pdfminer release
35
+
* Integrated new read layout OCR engine
36
+
37
+
**Bug fixes**
38
+
39
+
* Bug fix: Public version is no longer hard coded into the route
40
+
* Bug fix: Slow blob listings no longer block training
41
+
* Bug fix: Kestrel unable to process malformed JSON for Analyze requests
42
+
* Bug fix: Get/Delete APIs return errors for non-existent models
43
+
* Bug fix: Isolate currency characters not detected
44
+
* Bug fix: Prediction takes longer than 60 seconds
45
+
* Bug fix: Unable to extract key/value pairs. List index out of range.
46
+
47
+
### Form Recognizer Sample Labeling Tool is now open-source
48
+
49
+
The Form Recognizer Sample Labeling Tool is now available as an open-source project. You can integrate it within your solutions and make customer-specific changes to meet your needs.
50
+
51
+
For more information about the Form Recognizer Sample Labeling Tool, review the documentation available on [GitHub](https://github.com/microsoft/OCR-Form-Tools/blob/master/README.md).
52
+
53
+
### Labeling value types
54
+
55
+
Value types are now available for use with the Form Recognizer Sample Labeling Tool. These value types are currently supported:
56
+
57
+
* String
58
+
* Number
59
+
* Integer
60
+
* Date
61
+
* Time
62
+
63
+
This image shows what value type selection looks like within the Form Recognizer Sample Labeling Tool:
64
+
65
+
> [!div class="mx-imgBorder"]
66
+
> 
67
+
68
+
The extracted table are available in the JSON output in `pageResults`.
69
+
70
+
### Table visualization
71
+
72
+
The Form Recognizer Labeling Tool now displays tables that were recognized in the document. This lets you view the tables that have been recognized and extracted from the document, prior to labeling and analyzing with the Form Recognizer Sample Labeling Tool. This feature can be toggled on/off using the layers option.
73
+
74
+
This is an example of how tables are recognized and extracted:
75
+
76
+
> [!div class="mx-imgBorder"]
77
+
> 
78
+
79
+
> [!IMPORTANT]
80
+
> Labeling tables isn't supported. If tables are not recognized and extrated automatically, you can only label them as key/value pairs. When labeling tables as key/value pairs, please label each cell as a value.
81
+
82
+
## January 2020
83
+
84
+
This release introduces the Form Recognizer 2.0 (preview). In the sections below, you'll find more information about new features, enhancements, and changes.
0 commit comments