DetectTextInImage Operator

Description

The DetectTextInImage operator extracts text from images using Tesseract OCR (Optical Character Recognition). It supports multiple languages including English, Hindi, Tamil, and Telugu. The operator processes image files and returns the detected text as a string.

Model Information

  • OCR Engine: Tesseract OCR

  • Source: Google, via pytesseract Python wrapper

  • Supported Languages: English (eng), Hindi (hin), Tamil (tam), Telugu (tel)

  • Usage: The operator uses Tesseract’s OCR capabilities to extract text from images, enabling downstream tasks such as text analysis, content moderation, and document processing.

System Dependencies

  • Tesseract OCR

    • On Windows:

      1. Download from UB-Mannheim’s Tesseract

      2. Install and add to PATH

      3. Install language packs for Hindi, Tamil, and Telugu

    • On Linux: sudo apt install tesseract-ocr tesseract-ocr-hin tesseract-ocr-tam tesseract-ocr-tel

    • On macOS: brew install tesseract tesseract-lang

Operator Dependencies

  • pytesseract >= 0.3.10

  • Pillow >= 11.1.0

How to Run the Tests

  1. Ensure that you are in the root directory of the feluda project.

  2. Install dependencies (in your virtual environment):

    uv pip install "./operators/detect_text_in_image"
uv pip install "feluda[dev]"

  3. Ensure Tesseract OCR is installed and available in your PATH.

  4. Run the tests:

    pytest operators/detect_text_in_image/test.py

Usage

from feluda.factory import ImageFactory
from feluda.operators import DetectTextInImage

# Initialize the operator
operator = DetectTextInImage()

# Load an image
image = ImageFactory.make_from_url_to_path("https://example.com/image.png")

# Extract text
text = operator.run(image, remove_after_processing=False)
print(text)

# Check operator state
state = operator.state()
print(f"PSM: {state['psm']}, OEM: {state['oem']}")

# Cleanup resources
operator.cleanup()

Configuration

The operator accepts two configuration parameters:

  • psm (int): Page segmentation mode (default: 6)

  • oem (int): OCR Engine mode (default: 1)

These can be set during initialization:

operator = DetectTextInImage(psm=8, oem=3)
class operators.detect_text_in_image.detect_text_in_image.DetectTextInImage(psm: int = 6, oem: int = 1, tesseract_cmd: str | None = None)[source]

Bases: Operator

Operator to detect text in images using Tesseract OCR.

__init__(psm: int = 6, oem: int = 1, tesseract_cmd: str | None = None) → None[source]

Initialize the DetectTextInImage class.

Parameters:

  • psm (int) – Page segmentation mode for Tesseract (default: 6)

  • oem (int) – OCR Engine mode for Tesseract (default: 1)

validate_system() → None[source]

Validate that Tesseract OCR is installed and accessible.

Raises:

RuntimeError – If Tesseract is not installed or not in PATH.

validate_languages() → None[source]

Validate that required language packs are installed.

Checks for English, Hindi, Tamil, and Telugu language support.

run(file: ImageFactory, remove_after_processing: bool = False) → str[source]

Run the text detection operator.

Parameters:

  • file (ImageFactory) – ImageFactory object

  • remove_after_processing (bool) – Whether to remove the file after processing

Returns:

Detected text from the image

Return type:

str

cleanup() → None[source]

Cleans up resources used by the operator.

state() → dict[str, Any][source]

Returns the current state of the operator.

Returns:

State of the operator including PSM and OEM settings

Return type:

dict