Weakly-Supervised Cell Tracking via Backward-and-Forward Propagation

TL;DR

This paper introduces a weakly-supervised cell tracking approach that trains CNNs using only cell detection annotations, leveraging a backward-and-forward propagation method to infer cell associations without explicit association labels.

Contribution

The paper presents a novel backward-and-forward propagation technique that enables cell association inference from weakly-supervised detection CNNs, achieving performance comparable to fully supervised methods.

Findings

The method effectively associates cells using only detection annotations.

Performance nearly matches state-of-the-art supervised methods.

Code is publicly available for reproducibility.

Abstract

We propose a weakly-supervised cell tracking method that can train a convolutional neural network (CNN) by using only the annotation of "cell detection" (i.e., the coordinates of cell positions) without association information, in which cell positions can be easily obtained by nuclear staining. First, we train co-detection CNN that detects cells in successive frames by using weak-labels. Our key assumption is that co-detection CNN implicitly learns association in addition to detection. To obtain the association, we propose a backward-and-forward propagation method that analyzes the correspondence of cell positions in the outputs of co-detection CNN. Experiments demonstrated that the proposed method can associate cells by analyzing co-detection CNN. Even though the method uses only weak supervision, the performance of our method was almost the same as the state-of-the-art supervised method. Code is publicly available in https://github.com/naivete5656/WSCTBFP