Human3.6M: Large Scale Datasets and Predictive Methods for 3D Human Sensing in Natural Environments

• Published in: IEEE transactions on pattern analysis and machine intelligence Vol. 36; no. 7; pp. 1325 - 1339
• Main Authors: Ionescu, Catalin, Papava, Dragos, Olaru, Vlad, Sminchisescu, Cristian
• Format: Journal Article
• Language: English
• Published: Los Alamitos, CA IEEE 01.07.2014; IEEE Computer Society; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: 3D human pose estimation; Algorithms; Applied sciences; articulated body modeling; Artificial intelligence; Biometry - methods; Cameras; Computer science; control theory; systems; Computer systems and distributed systems. User interface; Databases, Factual; Ecosystem; Estimation; Exact sciences and technology; Fourier kernel approximations; human motion capture data; Humans; Image Enhancement - methods; Image Interpretation, Computer-Assisted - methods; Imaging, Three-Dimensional - methods; Information Storage and Retrieval - methods; Information systems. Data bases; Joints; large scale learning; Learning and adaptive systems; Matematik; Mathematical Sciences; Mathematics; Memory organisation. Data processing; Modeling and recovery of physical attributes; Motion; Natural Sciences; Naturvetenskap; optimization; Pattern Recognition, Automated - methods; Pattern recognition. Digital image processing. Computational geometry; Photography - methods; Posture; Reproducibility of Results; Sensitivity and Specificity; Sensors; Software; Solid modeling; structured prediction; Subtraction Technique; Three-dimensional displays; Training; Whole Body Imaging - methods; large-scale learning; structured prediction; articulated body modeling; optimization; human motion capture data; 3D human pose estimation; Fourier kernel approximations; Occlusion; Very large databases; Modeling; Optimization; Posture; Natural environment; Scientific research; Classification; Internal structure; Data structure; Learning algorithm; Computer vision; Statistical analysis; Motion estimation; Viewing angle; Computer animation; Augmented reality; Educational software program; Human activity; Time of flight method; Mixed reality; Data visualization; Body movement; Large scale; Occultation; Structural analysis

We introduce a new dataset, Human3.6M, of 3.6 Million accurate 3D Human poses, acquired by recording the performance of 5 female and 6 male subjects, under 4 different viewpoints, for training realistic human sensing systems and for evaluating the next generation of human pose estimation models and...