Abstract:Recent advances in 3D depth sensors have created many opportunities for security, surveillance, and entertainment. The 3D depth sensors provide more powerful monitoring systems for dangerous situations irrespective of lighting conditions in buildings or production facilities. To robustly recognize emergency actions or hazardous situations of workers at a production facility, we present human joint estimation and behavior recognition algorithms that solely use depth information in this paper. To estimate human joints on a low cost computing platform, we propose a human joint estimation algorithm that integrates a geodesic graph and a support vector machine (SVM). The human feature points are extracted within a range of geodesic distance from a geodesic graph. The geodesic graph is used for optimizing the estimation result. The SVM-based human joint estimator uses randomly selected human features to reduce computation. Body parts that typically involve many motions are then estimated by the geodesic distance value. The proposed algorithm can work for any human without calibration, and thus the system can be used with any subject immediately even with a low cost computing platform. In the case of the behavior recognition algorithm, the algorithm should have a simple behavior registration process, and it also should be robust to environmental changes. To meet these goals, we propose a template matching-based behavior recognition algorithm. Our method creates a behavior template set that consists of weighted human joint data with scale and rotation invariant properties. A single behavior template consists of the joint information that is estimated per frame. Additionally, we propose adaptive template rejection and a sliding window filter to prevent misrecognition between similar behaviors. The human joint estimation and behavior recognition algorithms are evaluated individually through several experiments and the performance is proven through a comparison with other algorithms. The experimental results show that our method performs well and is applicable in real environments.

Using Dynamic Programming To Match Human Behavior Sequences

Deep Behavior Tracing with Multi-level Temporality Preserved Embedding

A novel method for automated human behavior segmentation

Action Recognition Based on Global Optimal Similarity Measuring

Matching Model of Dance Movements and Music Rhythm Features Using Human Posture Estimation

A new approach to human motion sequence recognition with application to diving actions

Gait modeling and identifying based on dynamic template matching

Motion Recognition Based on Deep Learning and Human Joint Points

Human action recognition for 3D video based on action standard sequence

Human Action Recognition by Radon Transform

Recognition of Human Daily Action Based on Template Matching

Dynamic gesture retrieval: searching videos by human pose sequence

Human Activity Recognition with Posture Tendency Descriptors on Action Snippets

Weighted joint-based human behavior recognition algorithm using only depth information for low-cost intelligent video-surveillance system

Dynamic Human Behavior Pattern Detection and Classification

Human Action Recognition Based on Kinematic Similarity in Real Time.

Beyond Matching: Modeling Two-Sided Multi-Behavioral Sequences for Dynamic Person-Job Fit

Gait Recognition Via Flexible Measurement

Human Activity Recognition based on Dynamic Spatio-Temporal Relations

Extracting Key Postures in a Human Action Video Sequence.

Recognizing Human Activities by Key Frame in Video Sequences.