Cell detection and tracking are paramount for bio-analysis. Recent approachesrely on the tracking-by-model evolution paradigm, which usually consists oftraining end-to-end deep learning models to detect and track the cells on theframes with promising results. However, such methods require extensive amountsof annotated data, which is time-consuming to obtain and often requiresspecialized annotators. This work proposes a new approach based on theclassical tracking-by-detection paradigm that alleviates the requirement ofannotated data. More precisely, it approximates the cell shapes as orientedellipses and then uses generic-purpose oriented object detectors to identifythe cells in each frame. We then rely on a global data association algorithmthat explores temporal cell similarity using probability distance metrics,considering that the ellipses relate to two-dimensional Gaussian distributions.Our results show that our method can achieve detection and tracking resultscompetitively with state-of-the-art techniques that require considerably moreextensive data annotation. Our code is available at:https://github.com/LucasKirsten/Deep-Cell-Tracking-EBB.