Robust detection of heartbeats using association models from blood pressure and EEG signals.

BACKGROUNDS: The heartbeat is fundamental cardiac activity which is straightforwardly detected with a variety of measurement techniques for analyzing physiological signals. Unfortunately, unexpected noise or contaminated signals can distort or cut out electrocardiogram (ECG) signals in practice, misleading the heartbeat detectors to report a false heart rate or suspend itself for a considerable length of time in the worst case. To deal with the problem of unreliable heartbeat detection, PhysioNet/CinC suggests a challenge in 2014 for developing robust heart beat detectors using multimodal signals. METHODS: This article proposes a multimodal data association method that supplements ECG as a primary input signal with blood pressure (BP) and electroencephalogram (EEG) as complementary input signals when input signals are unreliable. If the current signal quality index (SQI) qualifies ECG as a reliable input signal, our method applies QRS detection to ECG and reports heartbeats. Otherwise, the current SQI selects the best supplementary input signal between BP and EEG after evaluating the current SQI of BP. When BP is chosen as a supplementary input signal, our association model between ECG and BP enables us to compute their regular intervals, detect characteristics BP signals, and estimate the locations of the heartbeat. When both ECG and BP are not qualified, our fusion method resorts to the association model between ECG and EEG that allows us to apply an adaptive filter to ECG and EEG, extract the QRS candidates, and report heartbeats. RESULTS: The proposed method achieved an overall score of 86.26 % for the test data when the input signals are unreliable. Our method outperformed the traditional method, which achieved 79.28 % using QRS detector and BP detector from PhysioNet. Our multimodal signal processing method outperforms the conventional unimodal method of taking ECG signals alone for both training and test data sets. CONCLUSIONS: To detect the heartbeat robustly, we have proposed a novel multimodal data association method of supplementing ECG with a variety of physiological signals and accounting for the patient-specific lag between different pulsatile signals and ECG. Multimodal signal detectors and data-fusion approaches such as those proposed in this article can reduce false alarms and improve patient monitoring.

Weakly Supervised Contrastive Learning for Unsupervised Vehicle Reidentification.

Reidentification (Re-id) of vehicles in a multicamera system is an essential process for traffic control automation. Previously, there have been efforts to reidentify vehicles based on shots of images with identity (id) labels, where the model training relies on the quality and quantity of the labels. However, labeling vehicle ids is a labor-intensive procedure. Instead of relying on expensive labels, we propose to exploit camera and tracklet ids that are automatically obtainable during a Re-id dataset construction. In this article, we present weakly supervised contrastive learning (WSCL) and domain adaptation (DA) techniques using camera and tracklet ids for unsupervised vehicle Re-id. We define each camera id as a subdomain and tracklet id as a label of a vehicle within each subdomain, i.e., weak label in the Re-id scenario. Within each subdomain, contrastive learning using tracklet ids is applied to learn a representation of vehicles. Then, DA is performed to match vehicle ids across the subdomains. We demonstrate the effectiveness of our method for unsupervised vehicle Re-id using various benchmarks. Experimental results show that the proposed method outperforms the recent state-of-the-art unsupervised Re-id methods. The source code is publicly available on https://github.com/andreYoo/WSCL_VeReid.

An Iterative Method for Unsupervised Robust Anomaly Detection Under Data Contamination.

Most deep anomaly detection models are based on learning normality from datasets due to the difficulty of defining abnormality by its diverse and inconsistent nature. Therefore, it has been a common practice to learn normality under the assumption that anomalous data are absent in a training dataset, which we call normality assumption. However, in practice, the normality assumption is often violated due to the nature of real data distributions that includes anomalous tails, i.e., a contaminated dataset. Thereby, the gap between the assumption and actual training data affects detrimentally in learning of an anomaly detection model. In this work, we propose a learning framework to reduce this gap and achieve better normality representation. Our key idea is to identify sample-wise normality and utilize it as an importance weight, which is updated iteratively during the training. Our framework is designed to be model-agnostic and hyperparameter insensitive so that it applies to a wide range of existing methods without careful parameter tuning. We apply our framework to three different representative approaches of deep anomaly detection that are classified into one-class classification-, probabilistic model-, and reconstruction-based approaches. In addition, we address the importance of a termination condition for iterative methods and propose a termination criterion inspired by the anomaly detection objective. We validate that our framework improves the robustness of the anomaly detection models under different levels of contamination ratios on five anomaly detection benchmark datasets and two image datasets. On various contaminated datasets, our framework improves the performance of three representative anomaly detection methods, measured by area under the ROC curve.

Abnormal Event Detection and Localization via Adversarial Event Prediction.

We present adversarial event prediction (AEP), a novel approach to detecting abnormal events through an event prediction setting. Given normal event samples, AEP derives the prediction model, which can discover the correlation between the present and future of events in the training step. In obtaining the prediction model, we propose adversarial learning for the past and future of events. The proposed adversarial learning enforces AEP to learn the representation for predicting future events and restricts the representation learning for the past of events. By exploiting the proposed adversarial learning, AEP can produce the discriminative model to detect an anomaly of events without complementary information, such as optical flow and explicit abnormal event samples in the training step. We demonstrate the efficiency of AEP for detecting anomalies of events using the UCSD-Ped, CUHK Avenue, Subway, and UCF-Crime data sets. Experiments include the performance analysis depending on hyperparameter settings and the comparison with existing state-of-the-art methods. The experimental results show that the proposed adversarial learning can assist in deriving a better model for normal events on AEP, and AEP trained by the proposed adversarial learning can surpass the existing state-of-the-art methods.