This paper presents a method to recognize human actions from sequences of depth maps. Specifically, we employ an action graph to model explicitly the dynamics of the actions and a bag of 3D points to characterize a set of salient postures that correspond to the nodes in the action graph. In addition, we propose a simple, but effective projection based sampling scheme to sample the bag of 3D points from the depth maps. Experimental results have shown that over 90% recognition accuracy were achieved by sampling only about 1% 3D points from the depth maps. Compared to the 2D silhouette based recognition, the recognition errors were halved. In addition, we demonstrate the potential of the bag of points posture model to deal with occlusions through simulation.

We present a new descriptor for activity recognition from videos acquired by a depth sensor. Previous descriptors mostly compute shape and motion features independently; thus, they often fail to capture the complex joint shapemotion cues at pixel-level. In contrast, we describe the depth sequence using a histogram capturing the distribution of the surface normal orientation in the 4D space of time, depth, and spatial coordinates. To build the histogram, we create 4D projectors, which quantize the 4D space and represent the possible directions for the 4D normal. We initialize the projectors using the vertices of a regular polychoron. Consequently, we refine the projectors using a discriminative density measure, such that additional projectors are induced in the directions where the 4D normals are more dense and discriminative. Through extensive experiments, we demonstrate that our descriptor better captures the joint shape-motion cues in the depth sequence, and thus outperforms the state-of-the-art on all relevant benchmarks.

Modern machine learning suffers from catastrophic forgetting when learning new classes incrementally. The performance dramatically degrades due to the missing data of old classes. Incremental learning methods have been proposed to retain the knowledge acquired from the old classes, by using knowledge distilling and keeping a few exemplars from the old classes. However, these methods struggle to scale up to a large number of classes. We believe this is because of the combination of two factors: (a) the data imbalance between the old and new classes, and (b) the increasing number of visually similar classes. Distinguishing between an increasing number of visually similar classes is particularly challenging, when the training data is unbalanced. We propose a simple and effective method to address this data imbalance issue. We found that the last fully connected layer has a strong bias towards the new classes, and this bias can be corrected by a linear model. With two bias parameters, our method performs remarkably well on two large datasets: ImageNet (1000 classes) and MS-Celeb-1M (10000 classes), outperforming the state-of-the-art algorithms by 11.1% and 13.2% respectively.

Ensembl Genomes (http://www.ensemblgenomes.org) is an integrating resource for genome-scale data from non-vertebrate species, complementing the resources for vertebrate genomics developed in the Ensembl project (http://www.ensembl.org). Together, the two resources provide a consistent set of programmatic and interactive interfaces to a rich range of data including genome sequence, gene models, transcript sequence, genetic variation, and comparative analysis. This paper provides an update to the previous publications about the resource, with a focus on recent developments and expansions. These include the incorporation of almost 20 000 additional genome sequences and over 35 000 tracks of RNA-Seq data, which have been aligned to genomic sequence and made available for visualization. Other advances since 2015 include the release of the database in Resource Description Framework (RDF) format, a large increase in community-derived curation, a new high-performance protein sequence search, additional cross-references, improved annotation of non-protein-coding genes, and the launch of pre-release and archival sites. Collectively, these changes are part of a continuing response to the increasing quantity of publicly-available genome-scale data, and the consequent need to archive, integrate, annotate and disseminate these using automated, scalable methods.

No abstract

Human action recognition is an important yet challenging task. Human actions usually involve human-object interactions, highly articulated motions, high intra-class variations, and complicated temporal structures. The recently developed commodity depth sensors open up new possibilities of dealing with this problem by providing 3D depth data of the scene. This information not only facilitates a rather powerful human motion capturing technique, but also makes it possible to efficiently model human-object interactions and intra-class variations. In this paper, we propose to characterize the human actions with a novel actionlet ensemble model, which represents the interaction of a subset of human joints. The proposed model is robust to noise, invariant to translational and temporal misalignment, and capable of characterizing both the human motion and the human-object interactions. We evaluate the proposed approach on three challenging action recognition datasets captured by Kinect devices, a multiview action recognition dataset captured with Kinect device, and a dataset captured by a motion capture system. The experimental evaluations show that the proposed approach achieves superior performance to the state-of-the-art algorithms.

We study the problem of action recognition from depth sequences captured by depth cameras, where noise and occlusion are common problems because they are captured with a single commodity camera. In order to deal with these issues, we extract semi-local features called random occupancy pattern (ROP) features, which employ a novel sampling scheme that effectively explores an extremely large sampling space. We also utilize a sparse coding approach to robustly encode these features. The proposed approach does not require careful parameter tuning. Its training is very fast due to the use of the high-dimensional integral image, and it is robust to the occlusions. Our technique is evaluated on two datasets captured by commodity depth cameras: an action dataset and a hand gesture dataset. Our classification results are superior to those obtained by the state of the art approaches on both datasets.

No abstract