Multi-label classification improves blood cell and aggregate identification
Multi-label versus multi-class classification of blood cells and their aggregates in microfluidic channels
Computer Vision and Pattern RecognitionMachine Learning
Summary
Blood analyzers often struggle to correctly identify groups of cells called aggregates. The authors compared two ways to classify images of blood cells: one that assigns a single cell type per image and another that allows multiple labels for cell aggregates. Their multi-label method can recognize cell groups even if the training data didn’t include those specific combinations. This approach reduces the need for detailed labeling and could improve clinical analysis of blood samples.
What this means in practice
- •For clinical laboratory technicians: Enhance automated blood analyzers to better classify cell aggregates without exhaustive labeling of training data.
- •For medical device developers: Develop improved blood analysis devices that incorporate multi-label classification to identify complex cell structures accurately.$Commercial implications: Enables more accurate blood diagnostics devices that address current limitations in aggregate cell analysis for clinical use.
Authors
Igor Zingman, Shada Abuhattum, Sara Kaliman, Maximilian Schlögel, Paul Müller, Markéta Kubánková, Nadine Ströhlein, Manuela Hauke, Lena Schnörer, Martin Kräter, Jochen Guck
Abstract
Deformability cytometry (DC) is a type of imaging flow cytometry, which uses a camera-equipped device to measure cellular stiffness in addition to other cellular properties at high throughput. Cellular properties such as area and elongation can identify cell types, but this requires prior knowledge of distinguishing properties and cannot be applied to clinically important cell aggregates. Using DC data, we evaluated conventional multi-class (MC) classification and introduced a multi-label (ML) approach for identifying blood cells and their aggregates. In particular, an ML classifier can simultaneously assign multiple cell-type labels to a single imaged event. We show that, unlike MC classification, ML classification can identify cell aggregates not represented in the training data. It also avoids the need for exhaustive, strictly defined aggregate labels, thereby simplifying and speeding up annotation. Since automated blood analyzers do not reliably analyze cell aggregates, our approach may help address this clinical gap.