Inquiry Regarding Google Summer of Code 2025 - Integrating LightGlue and ALIKED in OpenCV
محمد ايمن محمد
Dear OpenCV Team,
My name is Mohamed Kasm and I am a highly motivated student with a strong background in Computer Vision and experience with OpenCV. I am very interested in the "LightGlue Matcher with ALIKED Feature" project listed for Google Summer of Code 2025.
The prospect of enhancing OpenCV's feature matching capabilities by integrating the learning-based LightGlue matcher and the efficient ALIKED feature detector is particularly exciting to me. My experience in computer vision, including projects like "Computer Vision-Based Vehicle Detection" (as mentioned in my CV), along with my familiarity with Python, C++, and PyTorch, aligns well with the requirements of this project.
I have been reviewing the descriptions of LightGlue and ALIKED, and I would be grateful for the opportunity to discuss your team's specific vision and priorities for this integration. I am particularly interested in understanding:
- Are there specific parts of the OpenCV codebase that you would like me to focus on for this integration within the cv::Feature2D framework?
- Are there any specific considerations or challenges you foresee regarding the use of cv::dnn for integrating the PyTorch-based LightGlue model?
- Do you have any specific requirements or preferences for the implementation of the ALIKED feature detector within the cv::Feature2D interface?
- Are there any existing feature matching modules within OpenCV that you would recommend as a reference point for this integration?
I have also developed a preliminary proposal outlining my approach to this project, which I would be happy to share with you for your feedback and insights. I am very open to your suggestions and believe that a collaborative approach will lead to a successful integration that benefits the OpenCV community.
Thank you for your time and consideration. I look forward to hearing from you soon.
Sincerely,
Mohamed Kasm