Deep learning based domain adaptation for mitochondria segmentation on EM volumes.

BACKGROUND AND OBJECTIVE: Accurate segmentation of electron microscopy (EM) volumes of the brain is essential to characterize neuronal structures at a cell or organelle level. While supervised deep learning methods have led to major breakthroughs in that direction during the past years, they usually require large amounts of annotated data to be trained, and perform poorly on other data acquired under similar experimental and imaging conditions. This is a problem known as domain adaptation, since models that learned from a sample distribution (or source domain) struggle to maintain their performance on samples extracted from a different distribution or target domain. In this work, we address the complex case of deep learning based domain adaptation for mitochondria segmentation across EM datasets from different tissues and species. METHODS: We present three unsupervised domain adaptation strategies to improve mitochondria segmentation in the target domain based on (1) state-of-the-art style transfer between images of both domains; (2) self-supervised learning to pre-train a model using unlabeled source and target images, and then fine-tune it only with the source labels; and (3) multi-task neural network architectures trained end-to-end with both labeled and unlabeled images. Additionally, to ensure good generalization in our models, we propose a new training stopping criterion based on morphological priors obtained exclusively in the source domain. The code and its documentation are publicly available at https://github.com/danifranco/EM_domain_adaptation. RESULTS: We carried out all possible cross-dataset experiments using three publicly available EM datasets. We evaluated our proposed strategies and those of others based on the mitochondria semantic labels predicted on the target datasets. CONCLUSIONS: The methods introduced here outperform the baseline methods and compare favorably to the state of the art. In the absence of validation labels, monitoring our proposed morphology-based metric is an intuitive and effective way to stop the training process and select in average optimal models.

Optimization of Imminent Labor Prediction Systems in Women with Threatened Preterm Labor Based on Electrohysterography.

Preterm birth is the leading cause of death in newborns and the survivors are prone to health complications. Threatened preterm labor (TPL) is the most common cause of hospitalization in the second half of pregnancy. The current methods used in clinical practice to diagnose preterm labor, the Bishop score or cervical length, have high negative predictive values but not positive ones. In this work we analyzed the performance of computationally efficient classification algorithms, based on electrohysterographic recordings (EHG), such as random forest (RF), extreme learning machine (ELM) and K-nearest neighbors (KNN) for imminent labor (<7 days) prediction in women with TPL, using the 50th or 10th-90th percentiles of temporal, spectral and nonlinear EHG parameters with and without obstetric data inputs. Two criteria were assessed for the classifier design: F1-score and sensitivity. RFF1_2 and ELMF1_2 provided the highest F1-score values in the validation dataset, (88.17 ± 8.34% and 90.2 ± 4.43%) with the 50th percentile of EHG and obstetric inputs. ELMF1_2 outperformed RFF1_2 in sensitivity, being similar to those of ELMSens (sensitivity optimization). The 10th-90th percentiles did not provide a significant improvement over the 50th percentile. KNN performance was highly sensitive to the input dataset, with a high generalization capability.