i'm studying the Azure Custom Vision service for object detection, but I would like also to extrapolate text information within a tagged image zone.
Is it possible with Custom Vision?
If not, is it in the service roadmap?
Thank you
Related
I have implemented Azure Cognitive Read service to return extracted/OCR text from a PDF.
However, to make it easier for the user to understand the context/copy and paste data from the PDF i would like to overlay that text data over the PDF. I would then drop that PDF into a viewer.
Does anyone have any ideas on how to proceed. Would also be happy to use AWS. Basically an API that i submit a PDF to and it returns an OCR'd PDF is what i am after. If this is not possible, a library that i can submit the text and the PDF (and return a text searchable PDF) is also ideal.
I am looking for something similar and stumbled upon this:
https://learn.microsoft.com/en-us/azure/cognitive-services/form-recognizer/overview?tabs=v2-1
This is the Azure form recogniser.
What is Azure Form Recognizer?
Azure Form Recognizer is a part of Azure Applied AI Services that lets you build automated data processing software using machine learning technology. Identify and extract text, key/value pairs, selection marks, tables, and structure from your documents—the service outputs structured data that includes the relationships in the original file, bounding boxes, confidence and more. You quickly get accurate results that are tailored to your specific content without heavy manual intervention or extensive data science expertise. Use Form Recognizer to automate data entry in your applications and enrich your documents search capabilities.
They have an online example test:
https://fott-2-1.azurewebsites.net/prebuilts-analyze
Create a service in azure for free and test if it fits your needs. From there you will get a json reply and you can use the boundingBox to display. I haven't gone as far as applying the bounding box or something like that to the pdf.
You also have a free development nuget package to do exactly as you want:
https://ironsoftware.com/csharp/ocr/#sample-tesseract-create-searchable-pdf
OCR development is free for testing and see if it works for you.
I have a question regarding Azure Form Recognizer's OCR with handwritten text.
When running OCR on handwritten PDF files before labeling in Azure's Sample Labeling Tool, the OCR often detects text incorrectly. With other form analysis and extraction technologies, an option is often provided to enter the text that was supposed to be detected to essentially "correct" the OCR. For training Azure Form Recognizer in the Sample Labeling Tool (Docker image), I do not see a way for me to override the OCR text and enter the correct text.
Is there a way I can enter the text myself that the OCR is failing to detect or detecting incorrectly?
For example, the image below is what the OCR in Azure's Sample Labeling Tool picked up:
OCR detection sample image.
Is there a way to correct this result and tell Form Recognizer that the text should be: "Bridget Sims, MD"?
currently there there is no way to correct the OCR result and improve its accuracy right away. The typical scenario is to train a form recognizer model from a small set of training files, and use it to process more documents. During training, small amount of OCR errors are not essential to the model quality, you could ignore them. The product team is working on a new version of OCR with better handwriting recognition accuracy.
thanks
-xin
[Microsoft Azure Form Recognizer Team]
we want to create bicycle sharing project where google map is a must thing
we are willing to use flutter for that how ever it is totally depend on us which platform should be used.i am too much confused which platform should be used.
i have reseach and creating some sample apps in flutter and found that there is stability problems on flutter and integerating with google map is not up to the mark or not too much smart like uber (we want same look and feel like uber,ofo,uride).
please suggest which platform should be prefered.
Implementations like the rideshare companies you mention utilize several different Google Maps APIs, so it would go beyond just the google_maps_flutter Package functionality. A common example would be using Directions API[1] and Distance Matrix API, Web Service HTTP requests to generate routes and how long it'll take for all the nearby drivers to get to a user.
That's not to say it isn't possible with using multiple different other plugins (I've seen some for Google Maps Web Services for Flutter here[2]. It would be interesting to see how these applications perform in comparison to native Android; but I'm not sure if there's enough data to say yet.
[1] https://developers.google.com/maps/documentation/directions/intro
[2] https://pub.dartlang.org/packages/google_maps_webservice
I am working on a project for school and I need an OCR, I downloaded the free trial from abbyy cloud ocr sdk but after reading all the documentation and API I still don't understand how to use the cloud service. Did anyone have an experience using that tool and can explain to me how it works or send me a demo code for how to use it?
thanks!
Actually, dragging and dropping files onto the GUI should work.
Make sure that you have set the right recognition language. The default one is English.
If some short text fragments are not recognized and you don’t need to save any pictures and tables, you can try to use the textExtraction profile. The request URL should be like http://cloud.ocrsdk.com/processImage?profile=textExtraction&exportFormat=pdfSearchable.
If the document quality is poor, you can try to use field-level recognition (when text coordinates are specified directly). It could be tested using the same GUI sample (second tab).
Note that the recommended resolution for font size 12-16pt is about 300 dpi (more source image recommendations).
To get more detailed recommendations, you can send your images and Application ID to CloudOCRSDK#abbyy.com.
I'm currently working on a project which maps items to a physical location within a building. We have a map created in our GIS system (ArcGIS), and are able to resolve items to the appropriate shape ID on this map.
We want to create a web map using one of the APIs such as Google maps or Open Layers in order to display this info, directing the user to the correct location on the map, and ideally highlighting the particular location (for example using a shaded polygon), or at least identifying it with some sort of marker.
However we're very new to all of the online mapping APIs, and are finding it pretty overwhelming at the moment. All we really want is a flat map with our own custom imagery - we dont really care about geo-referencing, or projections etc.
Does anyone have any tips or info about how we might go about this? Even just general pointers would be helpful - it's hard to know where to look when there's so much info out there!
Cheers
If you are using ArcGIS Server, use ArcCatalog to expose your data as a WMS.
If you do not have an ArcGIS Server license, you can use MapServer or GeoServer to do the same (i.e expose as a WMS).
OpenLayers has support for adding WMS layers. See their examples section for sample code.
Cheers
OpenLayers can handle your own custom imagery maps without any problems.
You can check out this example:
http://openlayers.org/dev/examples/image-layer.html
And of course you need to read the API documentation
If you are using ArcGIS to creat eyour map, then use ArcGIS Server to serve up the maps online. Then use ArcGIS JavaScript API to build your web application. There are several out-of-the-box templates as well as easy-to-use examples for performing basic actions such as pan, zoom, select, highlight, redlining, etc.
Create your map service, then add your custom Map as a Dynamic Map Service or a Tiled Map Service (if you've cached it). Perform a Find (or Query if you have more than one feature you want to highlight). Add the Graphic from the FindResult to your GraphicsLayer.
Here's a more complete example.
You could also look at MapGuide OS and (if you don't have studio) Maestro. It is a server side application that will take in SHP files from Arc easily. The only downside IMO is that the built in search and reporting functionality is rather limited but can be extended via the API (C#, PHP and Java) that and it is rather centralized with a persistent process on the server (which can make debugging exceedingly fun).
The fusion/flexible (or whatever they heck they call it) layouts are based on the OpenLayers API but is still relatively new. And you can publish as a WMS or WFS I believe (I haven't done it personally).
If using ArcGIS Server, you can also expose your imagery and feature layers via a REST endpoint that can be consumed by OpenLayers. Check out James Fee's blog post on the topic.