GSoC/2016/StatusReports/OmarAmin

< GSoC‎ | 2016‎ | StatusReports
Jump to: navigation, search

DigiKam - Automatic Red Eyes detection and correction

digiKam includes a basic tool to fix red eyes in image editor. there is no automatic mode (detection and fix), which makes the process of fixing red eyes in a batch of images a time consuming task. The objective of this project is to add a feature to digiKam to detect and fix red eyes automatically, and include this tool in image editor and Batch Queue Manager.


Work Report

A Brief Description: The objective of this project is to add automatic red eye removal feature to digiKam BQM and Image Editor to detect and correct red eyes automatically.

Status: Following were the tasks for my GSoC'16 project:

  • Review current code from older Red Eyes Removal kipi tool which include some interesting algorithms.(Done)
  • Check all features required by the tool in all use cases (specially with BQM).(Done)
  • Implement low level filter for red eyes detection and correction(Done).
  • Implement GUI and integration into image editor.(In Progress)
  • Implement GUI and integration into Batch Queue Manager.(Done)
  • Write unit tests for 8 bits and 16 bits color depth images and different formats (TIFF, PBNG, JP2K, and PGF) and documentation.(In Progress)
  • Implement an automatic red eye correction filter for non-human eyes.(Future Plans)

Screenshot:

Redeye.JPG

Before Redeye Correction:

Before.JPG

After Redeye Correction in BQM:

After.JPG


IRC Nick: OmarAmin

IRC Channel: #digikam

This is awesome about GSoC: For the first time to be dealing with large code base, trying to write clean code and working on computer vision which is a field that i'm highly interested in.

This is what I learned during GSoC: Dealing with large code base, sharpening my C++ skills, working on a remote project with others, and improving my image processing and computer vision skills.


This page was last edited on 23 August 2016, at 15:05. Content is available under Creative Commons License SA 4.0 unless otherwise noted.