Update README.md

README.md

@@ -42,14 +42,19 @@ Run the service:
 
 - With GPU support:
 ```    
-docker run --rm --name pdf-document-layout-analysis --gpus '"device=0"' -p 5060:5060 --entrypoint ./start.sh huridocs/pdf-document-layout-analysis:v0.0.20
+docker run --rm --name pdf-document-layout-analysis --gpus '"device=0"' -p 5060:5060 --entrypoint ./start.sh huridocs/pdf-document-layout-analysis:v0.0.21
 ```
 
 - Without GPU support:
 ```
-docker run --rm --name pdf-document-layout-analysis -p 5060:5060 --entrypoint ./start.sh huridocs/pdf-document-layout-analysis:v0.0.20
+docker run --rm --name pdf-document-layout-analysis -p 5060:5060 --entrypoint ./start.sh huridocs/pdf-document-layout-analysis:v0.0.21
 ```
 
+[OPTIONAL] OCR the PDF. Check supported languages (curl localhost:5060/info):
+
+    curl -X POST -F 'language=en' -F 'file=@/PATH/TO/PDF/pdf_name.pdf' localhost:5060/ocr --output ocr_document.pdf
+
+
 Get the segments from a PDF:
 
     curl -X POST -F 'file=@/PATH/TO/PDF/pdf_name.pdf' localhost:5060
@@ -77,6 +82,12 @@ Start the service:
 
     make start
 
+
+[OPTIONAL] OCR the PDF. Check supported languages (curl localhost:5060/info):
+
+    curl -X POST -F 'language=en' -F 'file=@/PATH/TO/PDF/pdf_name.pdf' localhost:5060/ocr --output ocr_document.pdf
+
+
 Get the segments from a PDF:
 
     curl -X POST -F 'file=@/PATH/TO/PDF/pdf_name.pdf' localhost:5060
@@ -110,6 +121,9 @@ Even though the visual model using more resources than the others, generally it'
 "sees" the whole page and has an idea about all the context. On the other hand, LightGBM models are performing slightly worse
 but they are much faster and more resource-friendly. It will only require your CPU power.
 
+The service converts PDFs to text-searchable PDFs using [Tesseract OCR](https://github.com/tesseract-ocr/tesseract) and [ocrmypdf](https://ocrmypdf.readthedocs.io/en/latest/index.html).
+
+
 ## Data
 
 As we mentioned, we are using the visual model that trained on [DocLayNet](https://github.com/DS4SD/DocLayNet) dataset.