Recognizing human activities from video sequences or still images is a challenging task due to problems, such as background clutter, partial occlusion, changes in scale, viewpoint, lighting, and appearance. Many applications, including video surveillance systems, human-computer interaction, and robotics for human behavior characterization, require a multiple activity recognition system. In this work, we provide a detailed review of recent and state-of-the-art research advances in the field of human activity classification. We propose a categorization of human activity methodologies and discuss their advantages and limitations. In particular, we divide human activity classification methods into two large categories according to whether they use data from different modalities or not. Then, each of these categories is further analyzed into sub-categories, which reflect how they model human activities and what type of activities they are interested in. Moreover, we provide a comprehensive analysis of the existing, publicly available human activity classification datasets and examine the requirements for an ideal human activity recognition dataset. Finally, we report the characteristics of future research directions and present some open issues on human activity recognition.

Background Studies have demonstrated that the current US guidelines based on American College of Cardiology/American Heart Association (ACC/AHA) Pooled Cohort Equations Risk Calculator may underestimate risk of atherosclerotic cardiovascular disease ( CVD ) in certain high‐risk individuals, therefore missing opportunities for intensive therapy and preventing CVD events. Similarly, the guidelines may overestimate risk in low risk populations resulting in unnecessary statin therapy. We used Machine Learning ( ML ) to tackle this problem. Methods and Results We developed a ML Risk Calculator based on Support Vector Machines ( SVM s) using a 13‐year follow up data set from MESA (the Multi‐Ethnic Study of Atherosclerosis) of 6459 participants who were atherosclerotic CVD‐free at baseline. We provided identical input to both risk calculators and compared their performance. We then used the FLEMENGHO study (the Flemish Study of Environment, Genes and Health Outcomes) to validate the model in an external cohort. ACC / AHA Risk Calculator, based on 7.5% 10‐year risk threshold, recommended statin to 46.0%. Despite this high proportion, 23.8% of the 480 “Hard CVD ” events occurred in those not recommended statin, resulting in sensitivity 0.76, specificity 0.56, and AUC 0.71. In contrast, ML Risk Calculator recommended only 11.4% to take statin, and only 14.4% of “Hard CVD ” events occurred in those not recommended statin, resulting in sensitivity 0.86, specificity 0.95, and AUC 0.92. Similar results were found for prediction of “All CVD ” events. Conclusions The ML Risk Calculator outperformed the ACC/AHA Risk Calculator by recommending less drug therapy, yet missing fewer events. Additional studies are underway to validate the ML model in other cohorts and to explore its ability in short‐term CVD risk prediction.

A learning-based framework for action representation and recognition relying on the description of an action by time series of optical flow motion features is presented. In the learning step, the motion curves representing each action are clustered using Gaussian mixture modeling (GMM). In the recognition step, the optical flow curves of a probe sequence are also clustered using a GMM, then each probe sequence is projected onto the training space and the probe curves are matched to the learned curves using a non-metric similarity function based on the longest common subsequence, which is robust to noise and provides an intuitive notion of similarity between curves. Alignment between the mean curves is performed using canonical time warping. Finally, the probe sequence is categorized to the learned action with the maximum similarity using a nearest neighbor classification scheme. We also present a variant of the method where the length of the time series is reduced by dimensionality reduction in both training and test phases, in order to smooth out the outliers, which are common in these type of sequences. Experimental results on KTH, UCF Sports and UCF YouTube action databases demonstrate the effectiveness of the proposed method.

A human behavior recognition method with an application to political speech videos is presented. We focus on modeling the behavior of a subject with a conditional random field (CRF). The unary terms of the CRF employ spatiotemporal features (i.e., HOG3D, STIP and LBP). The pairwise terms are based on kinematic features such as the velocity and the acceleration of the subject. As an exact solution to the maximization of the posterior probability of the labels is generally intractable, loopy belief propagation was employed as an approximate inference method. To evaluate the performance of the model, we also introduce a novel behavior dataset, which includes low resolution video sequences depicting different people speaking in the Greek parliament. The subjects of the Parliament dataset are labeled as friendly, aggressive or neutral depending on the intensity of their political speech. The discrimination between friendly and aggressive labels is not straightforward in political speeches as the subjects perform similar movements in both cases. Experimental results show that the model can reach high accuracy in this relatively difficult dataset.

Incorporating additional knowledge in the learning process can be beneficial for several computer vision and machine learning tasks. Whether privileged information originates from a source domain that is adapted to a target domain, or as additional features available at training time only, using such privileged (i.e., auxiliary) information is of high importance as it improves the recognition performance and generalization. However, both primary and privileged information are rarely derived from the same distribution, which poses an additional challenge to the recognition task. To address these challenges, we present a novel learning paradigm that leverages privileged information in a domain adaptation setup to perform visual recognition tasks. The proposed framework, named Adaptive SVM+, combines the advantages of both the learning using privileged information (LUPI) paradigm and the domain adaptation framework, which are naturally embedded in the objective function of a regular SVM. We demonstrate the effectiveness of our approach on the publicly available Animals with Attributes and INTERACT datasets and report state-of-the-art results in both of them.