Integrate Tess4J OCR into Spring Boot: Step‑by‑Step Guide

Background

With the rapid advancement of information technology, extracting text from images is increasingly used for data entry and automation. Tess4J, a Java JNA wrapper for the Tesseract OCR engine, provides a powerful interface for this purpose, and can be seamlessly integrated into a Spring Boot application.

Part 1: Environment Setup

Before starting, ensure you have the following:

JDK 1.8 or higher

Maven

Latest Spring Boot version

Tess4J version 4.x or higher

Part 2: Add Dependency

Include the Tess4J dependency in your pom.xml:

<dependencies>
    <dependency>
        <groupId>net.sourceforge.tess4j</groupId>
        <artifactId>tess4j</artifactId>
        <version>4.5.4</version>
    </dependency>
    <!-- other dependencies -->
</dependencies>

Make sure the versions match your development environment.

Part 3: Add Tessdata Language Pack

Download the language data files, for example from:

https://gitcode.com/tesseract-ocr/tessdata/tree/main

or a Baidu Cloud link (password: 8v8u).

Part 4: Create OCR Service Class

@Service
public class OcrService {
    public String recognizeText(File imageFile) throws TesseractException {
        Tesseract tesseract = new Tesseract();
        // Set path to tessdata if using non‑default language packs
        tesseract.setDatapath("<your_tessdata_path>");
        tesseract.setLanguage("chi_sim");
        return tesseract.doOCR(imageFile);
    }

    public String recognizeTextFromUrl(String imageUrl) throws Exception {
        URL url = new URL(imageUrl);
        InputStream in = url.openStream();
        Files.copy(in, Paths.get("downloaded.jpg"), StandardCopyOption.REPLACE_EXISTING);
        File imageFile = new File("downloaded.jpg");
        return recognizeText(imageFile);
    }
}

The recognizeText(File imageFile) method performs OCR on a local file, while recognizeTextFromUrl(String imageUrl) first downloads a remote image before processing.

Part 5: Build REST Controller

@RestController
@RequestMapping("/api/ocr")
public class OcrController {
    private final OcrService ocrService;

    public OcrController(OcrService ocrService) {
        this.ocrService = ocrService;
    }

    @PostMapping("/upload")
    public ResponseEntity<String> uploadImage(@RequestParam("file") MultipartFile file) {
        try {
            File convFile = new File(System.getProperty("java.io.tmpdir") + "/" + file.getOriginalFilename());
            file.transferTo(convFile);
            String result = ocrService.recognizeText(convFile);
            return ResponseEntity.ok(result);
        } catch (Exception e) {
            e.printStackTrace();
            return ResponseEntity.badRequest().body("Recognition error: " + e.getMessage());
        }
    }

    @GetMapping("/recognize-url")
    public ResponseEntity<String> recognizeFromUrl(@RequestParam("imageUrl") String imageUrl) {
        try {
            String result = ocrService.recognizeTextFromUrl(imageUrl);
            return ResponseEntity.ok(result);
        } catch (Exception e) {
            e.printStackTrace();
            return ResponseEntity.badRequest().body("URL recognition error: " + e.getMessage());
        }
    }
}

The controller provides two endpoints: /api/ocr/upload for local image uploads and /api/ocr/recognize-url for processing images from a URL.

Part 6: Testing

Local test result:

Remote test result:

Conclusion

Following these steps gives you a Spring Boot service capable of recognizing text from both local and remote images. Adjust configurations such as language packs as needed for multilingual scenarios. While OCR still has room for improvement, Tess4J provides a solid starting point.