Загрузите отсканированный PDF-документ с помощью Tesseract OCR (бесплатная библиотека OCR) и сохраните результат как документ DOCX, используя C# и .NET.

using System.IO;
using SautinSoft.Document;
using System;

namespace Example
{
    class Program
    {
        static void Main(string[] args)
        {
            // Get your free trial key here:   
            // https://sautinsoft.com/start-for-free/

            LoadScannedPdf();
        }

        /// <summary>
        /// Load a scanned PDF document with help of Tesseract OCR (free OCR library) and save the result as DOCX document.
        /// </summary>
        /// <remarks>
        /// Details: https://www.sautinsoft.com/products/document/help/net/developer-guide/ocr-load-scanned-pdf-using-tesseract-and-save-as-docx-net-csharp-vb.php
        /// </remarks>
        static void LoadScannedPdf()
        {
            // Here we'll load a scanned PDF document (perform OCR) containing a text on English, Russian and Vietnamese.
            // Next save the OCR result as a new DOCX document.

            // First steps:

            // 1. Download data files for English, Russian and Vietnamese languages.
            // Please download the files: eng.traineddata, rus.traineddata and vie.traineddata.
            // From here (good and fast): https://github.com/tesseract-ocr/tessdata_fast
            // or (best and slow): https://github.com/tesseract-ocr/tessdata_best

            // 2. Copy the files: eng.traineddata, rus.traineddata and vie.traineddata to
            // the folder "tessdata" in the Project root.

            // 3. Be sure that the folder "tessdata" also contains "pdf.ttf" file.

            // Let's start:
            string inpFile = @"..\..\..\scan.pdf";
            string outFile = "Result.docx";

            PdfLoadOptions lo = new PdfLoadOptions();
            lo.OCROptions.OCRMode = OCRMode.Enabled;
            // 'false' - Never load embedded fonts in PDF. Use the fonts with the same name installed at the system or similar by font metrics.
			// 'true' - Always load embedded fonts in PDF.
            lo.PreserveEmbeddedFonts = true;            

            // You can specify all Tesseract parameters inside the method PerformOCR.
            lo.OCROptions.Method = PerformOCRTesseract;
            DocumentCore dc = DocumentCore.Load(inpFile, lo);

            // Make all text visible after Tesseract OCR (change font color to Black).
            // The matter is that Tesseract returns OCR result PDF document with invisible text.
            // But with help of Document .Net, we can change the text color, 
            // char scaling and spacing to desired.
            foreach (Run r in dc.GetChildElements(true, ElementType.Run))
            {
                r.CharacterFormat.FontColor = SautinSoft.Document.Color.Black;
                r.CharacterFormat.Scaling = 100;
                r.CharacterFormat.Spacing = 0;
            }

            dc.Save(outFile);

            // Open the result for demonstration purposes.
            System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(outFile) { UseShellExecute = true });
        }
        public static byte[] PerformOCRTesseract(byte[] image)
        {
            // Specify that Tesseract use three 3 languages: English, Russian and Vietnamese.
            string tesseractLanguages = "eng+rus+vie";

            // A path to a folder which contains languages data files and font file "pdf.ttf".
            // Language data files can be found here:
            // Good and fast: https://github.com/tesseract-ocr/tessdata_fast
            // or
            // Best and slow: https://github.com/tesseract-ocr/tessdata_best
            // Also this folder must have write permissions.
            string tesseractData = Path.GetFullPath(@"..\..\..\tessdata\");

            // A path for a temporary PDF file (because Tesseract returns OCR result as PDF document)
            string tempFile = Path.Combine(tesseractData, Path.GetRandomFileName());            

            try
            {
                using (Tesseract.IResultRenderer renderer = Tesseract.PdfResultRenderer.CreatePdfRenderer(tempFile, tesseractData, true))
                {
                    using (renderer.BeginDocument("Serachablepdf"))
                    {
                        using (Tesseract.TesseractEngine engine = new Tesseract.TesseractEngine(tesseractData, tesseractLanguages, Tesseract.EngineMode.Default))
                        {
                            engine.DefaultPageSegMode = Tesseract.PageSegMode.Auto;
                            
                            byte[] imgBytes = image;
                            using (Tesseract.Pix img = Tesseract.Pix.LoadFromMemory(imgBytes))
                            {
                                using (var page = engine.Process(img, "Serachablepdf"))
                                {
                                    renderer.AddPage(page);
                                }
                            }
                        }
                    }
                }

                return File.ReadAllBytes(tempFile + ".pdf");
            }
            catch (Exception e)
            {
                Console.WriteLine();
                Console.WriteLine("Please be sure that you have Language data files (*.traineddata) in your folder \"tessdata\"");
                Console.WriteLine("The Language data files can be download from here: https://github.com/tesseract-ocr/tessdata_fast");
                Console.ReadKey();
                throw new Exception("Error Tesseract: " + e.Message);
            }
            finally
            {
               
            }
        }        
    }
}

Download