Scribd
Upload
4 views

Design Phase

Uploaded by

Gjwkw Bai
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
4 views

Design Phase

Uploaded by

Gjwkw Bai
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 10

Text Detection and

Extraction using
OpenCV and OCR
This presentation explores the process of extracting text from
images using OpenCV and OCR, highlighting key concepts,
algorithms, and techniques.

pk
by pavan kalyan
Introduction to OpenCV and OCR
OpenCV OCR

Open Source Computer Vision Library, a powerful Optical Character Recognition, a technology that
toolkit for image and video processing. It provides a converts images of text into machine-readable text.
wide range of functions for tasks like image OCR algorithms analyze image features to identify
manipulation, analysis, and object recognition. characters and reconstruct the original text.
Importance of Text
Detection and
Extraction
Document Digitization Image Recognition
Converting paper Enabling image search
documents to digital engines to understand and
formats for easier storage, retrieve relevant images
search, and analysis. based on text content.

Autonomous Systems
Assisting self-driving cars to interpret traffic signs and navigate
roads safely.
Architecture Overview
Image Preprocessing Character Recognition
Noise reduction, image enhancement, and Recognizing individual characters using OCR
resizing to optimize for OCR. algorithms to transform image data into text.

1 2 3 4

Text Detection Post-processing


Identifying text regions within the image, Correcting errors, cleaning up the output text,
separating them from non-textual content. and formatting for desired results.
Text Detection Algorithms
1 Contour-Based 2 EAST Text Detector
Detection
A deep learning-based
Finding connected shapes algorithm that directly
in the image and filtering predicts text regions and
for text-like contours. their bounding boxes.

3 MSER
Maximally Stable Extremal Regions, an algorithm based on
identifying stable regions within an image that are likely to
contain text.
Contour-Based Detection
Steps Advantages Limitations

Convert the image to grayscale, Simple to implement, relatively Can be sensitive to noise, less
threshold it to create a binary efficient, and works well for images accurate for complex text layouts,
image, find contours, and filter with clear text regions. and may struggle with curved or
based on shape and size. irregular text.
EAST Text Detector
Deep Learning
Uses a convolutional neural network (CNN) to learn
features and predict text regions directly from image
data.

Real-Time Performance
Optimized for speed, making it suitable for real-time
applications, such as video processing and live text
recognition.

High Accuracy
Achieves high accuracy, especially for complex text
layouts and challenging images.
Text Extraction Methods

OCR Engines
1 Specialized software libraries designed for text recognition from images.

Tesseract
2 Open source OCR engine, known for its high accuracy and
support for multiple languages.

EasyOCR
3 Python-based OCR library, offering simplicity, ease of
use, and integration with other libraries.
Pytesseract
1 2
Installation Image Input
Requires Tesseract engine to be Takes an image as input and
installed separately. performs OCR.

3
Text Output
Returns recognized text as a
string.
EastOCR

Fast Multi-Language
Designed for speed and Supports a wide range of
efficiency, suitable for real-time languages.
applications.

Simple API
Easy to use with a
straightforward API.

You might also like