We are looking for a dozen students to help us build a unique data product which will advance research in a number of domains, such as Space Weather, Heliophysics, and Computer Vision. This data product is the cornerstone of our multi-year project, which will consequently feed many other research efforts to come.

We are particularly interested in students who:

• take pride in the quality and precision of their work, (this is #1 reason for students to stay on the project) 

• feel very comfortable using their computer and mouse (to paint over objects exactly as they appear on the screen),

• are enrolled in a BS/MS program in GSU, preferably one from either the Physics & Astronomy department or the Computer Science department,

• have access to a computer device (with a mouse and webcam) and a reliable internet connection.

We prioritize:

• people with disability (this is our #1 priority due to lack of enough opportunities for this community that allow fully-remote and asynchronous participation), 

• racial/ethnic and gender/sexuality minority groups,

• those who currently do not have any other paid, on-campus responsibility.

• students from relevant programs (CS, IT, Physics)

Start Date: Immediately after an online training session (mandatory)

Work Duration: up to 14 weeks (depending on the quality of the participant's work)

Work Demand: 8 hours / week (asynchronous)

Compensation: $15/hour (i.e., $120 for finishing one week's load of work; there is no partial payment)

We want to train our algorithm to do what trained eyes can do for a specific task. This starts with gathering a relatively large collection of what humans visually perceive of our target object, i.e., solar filaments. To do so, we need to first train a team of students to annotate solar filaments and then record their visual perception of many examples. This is where your contribution comes into play. When your annotation task is done, the product will pave the road for an array of research topics that is of interest to scientists in domains such as Heliophysics, Space Weather, and Computer Vision. 

We teamed up with V7 for annotation of the H-Alpha images of the Sun. With an array of user-friendly tools offered by V7, not only can we speed up our manual annotation process by an order of magnitude but we can also produce much more precisely annotated data compared to what classical tools would have given us.

In the videos below (copied from V7's website), some of the tools available in V7 are surveyed.

The graphic below summarizes the flow of our work on a weekly basis. Our team will take care of many behind-the-scenes activities to ensure that the annotators receive the best quality of images, well-organized and on-time. As the diagram suggests, the entire team can move on to the next week if and only if each and every annotator finishes the annotation of their assigned images.

The team will be divided into sub-teams so that each image is annotated by multiple annotators, independently. The only team-work expected is the (bi-) weekly milestones.

One of the critical requirements in the annotation process is the precision of the annotated objects. For this reason, we avoid overloading the annotators with tasks. Instead, we expect scrupulous care and commitment to the quality of work.

This is remote work; the entire process is anticipated to be carried out online (including the training phase). Thus, responsiveness of the participants is highly important. We expect the annotators to be ready for ad hoc virtual meetings (with the camera-on policy).

The annotation task is asynchronous. That is, the annotators are free to work on their assignments at their desired time. That said, the (bi-) weekly milestones must be met. This is critical because the entire group can move forward to the next week's assignment if and only if every annotator has completed their task.