Line-level text detection with Surya

In this tutorial we will perform line-level text detection using Surya toolkit and OpenVINO.

*image source

Model used for line-level text detection based on Segformer. It has the following features:

• It is specialized for document OCR. It will likely not work on photos or other images.
• It is for printed text, not handwriting.
• The model has trained itself to ignore advertisements.
• Languages with very different character sets may not work well.

Table of contents:

1. Fetch test image.
2. Run PyTorch inference.
3. Convert model to OpenVINO Intermediate Representation (IR) format.
4. Run OpenVINO model.
5. Apply post-training quantization using NNCF
   1. Prepare dataset
   2. Quantize model
6. Run quantized OpenVINO model
7. Interactive inference.

Installation instructions

This is a self-contained example that relies solely on its own code.
We recommend running the notebook in a virtual environment. You only need a Jupyter server to start. For details, please refer to Installation Guide.