Quran Page Line Segmentation Program
Budget: $10 – $30 USD
Task Title:
Line Segmentation from Quran Page Image (Arabic, Right-to-Left, with Diacritics)
? Task Description:
I have a scanned image of a Quran page in Arabic (PNG), and I want a program (in C# or Python) that performs accurate horizontal line segmentation.
Each line of Quranic text must be cropped into a separate image (PNG).
The code must:
Cut the image into exactly horizontal lines ( in the example they are 15 , they can be more in other images)
Detect ink pixels accurately — any pixel that is not white or transparent is considered "ink" (including colored pixels).
Detect line content precisely and separate any overflowing parts of text (like dots or diacritics that belong to the next line) intelligently.
? Requirements:
✅ The code must:
Work right-to-left (Arabic script)
Identify the true height of each line, and cut only full content
If any "ink" or text parts (like dots or tashkeel or symbols) are mistakenly captured at the bottom of a line while they belong to the next line, those "overflow bits" should be:
Moved to the next line (not deleted)
Without losing any visual pixels from the original image
Preserve the dimensions: The dimensions of the resulting image must equal the dimensions of the original page.
Result in 15 line images(or more depending on the page), each containing one full line only — no bits from other lines
Avoid aggressively trimming the lines, which may cut letters or words
? Testing Criteria:
To consider the task done correctly:
The resulting folder contains 15 images, one per line
Lines must not contain any leftover bits at the bottom that belong to next line
All words and diacritics should be complete
? Input:
A Quran page image ( .png)
Page is in Arabic, with decorative script and diacritics
? Output:
A folder containing: line_01.png, line_02.png, ..., line_15.png
Each is a clean horizontal line with complete content and no cross-line overlap
? Languages:
✅ Python
✅ Or C# (.NET Framework or .NET Core, using System.Drawing or similar)
Line Segmentation from Quran Page Image (Arabic, Right-to-Left, with Diacritics)
? Task Description:
I have a scanned image of a Quran page in Arabic (PNG), and I want a program (in C# or Python) that performs accurate horizontal line segmentation.
Each line of Quranic text must be cropped into a separate image (PNG).
The code must:
Cut the image into exactly horizontal lines ( in the example they are 15 , they can be more in other images)
Detect ink pixels accurately — any pixel that is not white or transparent is considered "ink" (including colored pixels).
Detect line content precisely and separate any overflowing parts of text (like dots or diacritics that belong to the next line) intelligently.
? Requirements:
✅ The code must:
Work right-to-left (Arabic script)
Identify the true height of each line, and cut only full content
If any "ink" or text parts (like dots or tashkeel or symbols) are mistakenly captured at the bottom of a line while they belong to the next line, those "overflow bits" should be:
Moved to the next line (not deleted)
Without losing any visual pixels from the original image
Preserve the dimensions: The dimensions of the resulting image must equal the dimensions of the original page.
Result in 15 line images(or more depending on the page), each containing one full line only — no bits from other lines
Avoid aggressively trimming the lines, which may cut letters or words
? Testing Criteria:
To consider the task done correctly:
The resulting folder contains 15 images, one per line
Lines must not contain any leftover bits at the bottom that belong to next line
All words and diacritics should be complete
? Input:
A Quran page image ( .png)
Page is in Arabic, with decorative script and diacritics
? Output:
A folder containing: line_01.png, line_02.png, ..., line_15.png
Each is a clean horizontal line with complete content and no cross-line overlap
? Languages:
✅ Python
✅ Or C# (.NET Framework or .NET Core, using System.Drawing or similar)