Collaboration call - photometric methods for visual tracking

I’m seeking one or more collaborators to help implement an open-source Julia implementation of a photometric method (sometimes also referred to as a direct method) for visual tracking, i.e. tracking objects or camera motion in video.

I developed and implemented this framework in C++ at a company I worked at previously. It was state of the art in terms of it’s accuracy and robustness, especially in the face of local brightness changes. It can be applied to rigid 3D object tracking, camera calibration, visual odometry, SLAM, and large scale scene reconstruction.

The method has been published, but no implementation has been released publicly (that I’m aware of) in any language. This project is therefore an opportunity for the Julia ecosystem to have the state of the art method, and attract more people to Julia.

I have already implemented the non-linear solver part of the framework. However, I’ve ground to a bit of a halt with the next steps. I would love to work with a motivated programmer to get the rest done. I can offer tuition on the method, and design and review of the implementation.

I’m happy to share details and results with anyone who has a genuine interest in collaborating, to give an idea of what the method is capable of.


I’d be very interested in this project, and would love to help as much as I can. I’m not sure how much I will be able to help however.

This is super exciting!

1 Like

I’m not sure how much of the methods/designs overlap, but @baggepinnen has BlobTracking.jl that appears to have video tracking capabilities.

1 Like

My guess would be very little overlap. Also, BlobTracking.jl hasn’t been touched in ages, so I encourage you to disregard it

1 Like

Hello, I am excited to be a part of this collaboration. Not sure how much I can offer but will be actively available to give a shot at it.

1 Like

Just a quick note to say that I appreciate all offers made :pray:, and have & will DM anyone wanting to collaborate.

Fwiw the annual “Summer of Code” program makes small amounts of funding available for project mentorships of this kind.