BiPOCO: Bi-Directional Trajectory Prediction with Pose Constraints for Pedestrian Anomaly Detection

Abstract We present BiPOCO, a Bi-directional trajectory predictor with POse COnstraints, for detecting anomalous activities of pedestrians in videos. In contrast to prior work based on feature reconstruction, our work identifies pedestrian anomalous events by forecasting their future trajectories and comparing the predictions with their expectations. We introduce a set of novel compositional pose-based losses …

Leveraging Trajectory Prediction for Pedestrian Video Anomaly Detection

Abstract Video anomaly detection is a core problem in vision. Correctly detecting and identifying anomalous behaviors in pedestrians from video data will enable safety-critical applications such as surveillance, activity monitoring, and human-robot interaction. In this paper, we propose to leverage trajectory localization and prediction for unsupervised pedestrian anomaly event detection. Different than previous reconstruction-based approaches, …

Angle-Regulated Transformer Network for Pedestrian Trajectory Prediction

Abstract Pedestrian trajectory prediction is an important topic in autonomous vehicle (AV) research. Accurate predictions of future pedestrian locations and movements are essential for AVs to safely and efficiently navigate in dense urban environments. The recently developed Transformer (TF) network has shown superior results for applications such as natural language processing and machine translation, but …

Multiple Instance Choquet Integral with Binary Fuzzy Measures for Remote Sensing Classifier Fusion with Imprecise Labels

Abstract Classifier fusion methods integrate complementary information from multiple classifiers or detectors and can aid remote sensing applications such as target detection and hyperspectral image analysis. The Choquet integral (CI), parameterized by fuzzy measures (FMs), has been widely used in the literature as an effective non-linear fusion framework. Standard supervised CI fusion algorithms often require …

Stochastic Sampling Simulation for Pedestrian Trajectory Prediction

Abstract Urban environments pose a significant challenge for autonomous vehicles (AVs) as they must safely navigate while in close proximity to many pedestrians. It is crucial for the AV to correctly understand and predict the future trajectories of pedestrians to avoid collision and plan a safe path. Deep neural networks (DNNs) have shown promising results …

Efficient Binary Fuzzy Measure Representation and Choquet Integral Learning

Abstract The Choquet integral (ChI), a parametric function for information aggregation, is parameterized by the fuzzy measure (FM), which has 2^N real-valued variables for N inputs. However, the ChI incurs huge storage and computational burden due to its exponential complexity relative to N and, as a result, its calculation, storage, and learning becomes intractable for …

Environmentally-Adaptive Target Recognition for SAS Imagery

Abstract Characteristics of underwater targets displayed in synthetic aperture sonar (SAS) imagery vary depending on their environmental context. Discriminative features in sea grass may differ from the features that are discriminative in sand ripple, for example. Environmentally-adaptive target detection and classification systems that take into account environmental context, therefore, have the potential for improved results. …

Multiple-instance Learning-based Sonar Image Classification

Abstract An approach to image labeling by seabed context based on multiple-instance learning via embedded instance selection (MILES) is presented. Sonar images are first segmented into superpixels with associated intensity and texture feature distributions. These superpixels are defined as the “instances” and the sonar images are defined as the “bags” within the MILES classification framework. …

Multiple Instance Choquet Integral for Classifier Fusion

Abstract The Multiple Instance Choquet integral (MICI) for classifier fusion and an evolutionary algorithm for parameter estimation is presented. The Choquet integral has a long history of providing an effective framework for non-linear fusion. Previous methods to learn an appropriate measure for the Choquet integral assumed accurate and precise training labels (with low levels of …