Colloidoscope: Detecting Dense Colloids in 3d with Deep Learning
Abdelwahab Kawafi, Lars K\"urten, Levke Ortlieb, Yushi Yang, Abraham, Mauleon Amieva, James E. Hallett, C.Patrick Royall

TL;DR
Colloidoscope is a deep learning pipeline that improves detection and tracking of dense colloids in 3D confocal microscopy images, outperforming traditional methods especially in challenging high-density and low-contrast conditions.
Contribution
This work introduces a 3D residual Unet-based deep learning method trained on simulated data to enhance colloid detection in complex imaging scenarios, demonstrating robustness and high accuracy.
Findings
Achieves higher recall than heuristic methods
Maintains high precision in particle detection
Robust to photobleaching effects
Abstract
Colloidoscope is a deep learning pipeline employing a 3D residual Unet architecture, designed to enhance the tracking of dense colloidal suspensions through confocal microscopy. This methodology uses a simulated training dataset that reflects a wide array of real-world imaging conditions, specifically targeting high colloid volume fraction and low-contrast scenarios where traditional detection methods struggle. Central to our approach is the use of experimental signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and point-spread-functions (PSFs) to accurately quantify and simulate the experimental data. Our findings reveal that Colloidoscope achieves superior recall in particle detection (finds more particles) compared to conventional heuristic methods. Simultaneously, high precision is maintained (high fraction of true positives.) The model demonstrates a notable robustness to…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsPhotoacoustic and Ultrasonic Imaging
