Interactive Cell Tracking

In this experiment, you can mark a cell at the beginning of a video and then automatically follow that cell through the rest of the video. The experiment uses an interactive segmentation neural network (Segment Anything) and uses it to follow the cell through time. Errors in tracking the cell can be corrected by re-tagging it.

How does tracking the cell over time work?

The cell is segmented by marking the pixel at the beginning of the video, as in the segmentation example. Then the outline of the cell identified in this way is transferred to the next frame of the video, and the segmentation model is used with this outline as a guide to find the corresponding cell again. This process is repeated until the cell is identified throughout the video. If there are errors in the result, they can be corrected by re-tagging the corresponding video segment.

Application areas of tracking

Automatic following of objects in videos has many applications, e.g. in self-driving cars to understand the movement of other road users. The methods used there are similar to the application shown here: following cells in microscopic time recordings. This technique is used in biological and medical research to test the behavior of cell cultures and their response to different influences, such as drug treatments. Experiments on cell cultures are a common precursor to drug studies in animals or patients.

---

Despite careful control of the content, we assume no liability for the content of external links. The operators of the linked pages are solely responsible for their content.