How can I detect lines in images or PDFs?

In today's world, an enormous amount of visual data is available in the form of images and PDF files. These visual files have become an integral part of our daily lives, from photos to scanned documents. However, to make sense of this visual data, it is essential to convert it into structured data that can be analyzed and processed. Most of the time, a person does this job. Can you imagine how much time is spent on a simple process?

As you know, computer vision is making enormous improvements every day. And the OCR technology helps you to create an autonomous process to convert visual data to structured. But the problem is finding the suitable OCR model in your use case and running it fastly. And the Cameralyze is here to help you to do that. We have lots of product-ready OCR models for you. All of them are serverless, which means you don't need to think to scale your structure, and also you can create end-to-end flow if you want.

Today, I want to talk more about the Line Detection model. That model is more than a standard OCR model. It detects words on the visual data and merges them to create a line. The model has a confidence configuration that helps you to control the minimum confidence threshold.

Okay, let's start to use it now. You can access the model here, and I also added a demo program below.

It's time to DIY. As a first step, you need to login and get your API Key. If you didn't register yet, click here.

import cameralyze

# connect to Cameralyze
connector = cameralyze.Model(api_key="YOUR_API_KEY")

# set the model you want to use
connector.set_model(model="1dfa8426-d559-41a1-a8cd-cfbdc936592e")

# run for image URL
image = "https://"

# run for local file
image = model.read_file(path="")

# run for base64 file
import base64
with open("yourfile.jpg", "rb") as image_file: 
  image = base64.b64encode(image_file.read()).decode("utf8")
  
  
# get response
response = model.predict(image=image)

The response of the model structure is:

{
    "confidence": "Float",
    "left": "Integer",
    "top": "Integer",
    "width": "Integer",
    "height": "Integer",
    "word": "String"
}