Federico Iuricich
@iuricichf.bsky.social
2/
We benchmark popular tracking methods and show where they fail under sparse sampling.
🔍 More displacement
🔍 Missed mitosis
🔍 Tracking ambiguity
We also outline a framework for evaluating trackers, and we discuss what’s next for robust tracking in these settings.
📄 t.co/rAiTPCb8KX
We benchmark popular tracking methods and show where they fail under sparse sampling.
🔍 More displacement
🔍 Missed mitosis
🔍 Tracking ambiguity
We also outline a framework for evaluating trackers, and we discuss what’s next for robust tracking in these settings.
📄 t.co/rAiTPCb8KX
June 12, 2025 at 10:22 PM
2/
We benchmark popular tracking methods and show where they fail under sparse sampling.
🔍 More displacement
🔍 Missed mitosis
🔍 Tracking ambiguity
We also outline a framework for evaluating trackers, and we discuss what’s next for robust tracking in these settings.
📄 t.co/rAiTPCb8KX
We benchmark popular tracking methods and show where they fail under sparse sampling.
🔍 More displacement
🔍 Missed mitosis
🔍 Tracking ambiguity
We also outline a framework for evaluating trackers, and we discuss what’s next for robust tracking in these settings.
📄 t.co/rAiTPCb8KX
1/ Most tracking methods assume high-frame-rate imaging, but that’s not always feasible:
⚠️ Photobleaching
⚠️ Phototoxicity
➡️ We propose a new dataset with long-term, sparsely sampled sequences, complete with ground-truth for ID, mitosis, and motion. [https://davislab.github.io/LFCT-dataset/]
⚠️ Photobleaching
⚠️ Phototoxicity
➡️ We propose a new dataset with long-term, sparsely sampled sequences, complete with ground-truth for ID, mitosis, and motion. [https://davislab.github.io/LFCT-dataset/]
June 12, 2025 at 10:22 PM
1/ Most tracking methods assume high-frame-rate imaging, but that’s not always feasible:
⚠️ Photobleaching
⚠️ Phototoxicity
➡️ We propose a new dataset with long-term, sparsely sampled sequences, complete with ground-truth for ID, mitosis, and motion. [https://davislab.github.io/LFCT-dataset/]
⚠️ Photobleaching
⚠️ Phototoxicity
➡️ We propose a new dataset with long-term, sparsely sampled sequences, complete with ground-truth for ID, mitosis, and motion. [https://davislab.github.io/LFCT-dataset/]