Track & Field


Publish your raw data and your speculations, then let other people do the analysis: track and field edition

#artificialintelligence

Methods 2127 observations of competition best performances and mass spectrometry-measured serum androgen concentrations, obtained during the 2011 and 2013 International Association of Athletics Federations World Championships, were analysed in male and female elite track and field athletes. To test the influence of serum androgen levels on performance, male and female athletes were classified in tertiles according to their free testosterone (fT) concentration and the best competition results achieved in the highest and lowest fT tertiles were then compared. Results The type of athletic event did not influence fT concentration among elite women, whereas male sprinters showed higher values for fT than male athletes in other events. Men involved in all throwing events showed significantly (p 0.05) lower testosterone and sex hormone binding globulin than men in other events. When compared with the lowest female fT tertile, women with the highest fT tertile performed significantly (p 0.05) better in 400 m, 400 m hurdles, 800 m, hammer throw, and pole vault with margins of 2.73%, 2.78%, 1.78%, 4.53%, and 2.94%, respectively. Such a pattern was not found in any of the male athletic events. I'm sure you wouldn't be surprised to see these kinds of mistakes in published work. What is more distressing is that this evidence is said to be a key submission in the IAAF's upcoming case against CAS [the Court of Arbitration for Sport], since the CAS has argued that sex classification on the basis of T levels are only justified if high T confers a "significant competitive advantage".


Soft Video Parsing by Label Distribution Learning

AAAI Conferences

In this paper, we tackle the problem of segmenting out a sequence of actions from videos. The videos contain background and actions which are usually composed of ordered sub-actions. We refer the sub-actions and the background as semantic units. Considering the possible overlap between two adjacent semantic units, we utilize label distributions to annotate the various segments in the video. The label distribution covers a certain number of semantic unit labels, representing the degree to which each label describes the video segment. The mapping from a video segment to its label distribution is then learned by a Label Distribution Learning (LDL) algorithm. Based on the LDL model, a soft video parsing method with segmental regular grammars is proposed to construct a tree structure for the video. Each leaf of the tree stands for a video clip of background or sub-action. The proposed method shows promising results on the THUMOS'14 and MSR-II datasets and its computational complexity is much less than the state-of-the-art method.


Submodular Attribute Selection for Action Recognition in Video

Neural Information Processing Systems

In real-world action recognition problems, low-level features cannot adequately characterize the rich spatial-temporal structures in action videos. In this work, we encode actions based on attributes that describes actions as high-level concepts: \textit{e.g.}, jump forward and motion in the air. We base our analysis on two types of action attributes. One type of action attributes is generated by humans. The second type is data-driven attributes, which is learned from data using dictionary learning methods. Attribute-based representation may exhibit high variance due to noisy and redundant attributes. We propose a discriminative and compact attribute-based representation by selecting a subset of discriminative attributes from a large attribute set. Three attribute selection criteria are proposed and formulated as a submodular optimization problem. A greedy optimization algorithm is presented and guaranteed to be at least (1-1/e)-approximation to the optimum. Experimental results on the Olympic Sports and UCF101 datasets demonstrate that the proposed attribute-based representation can significantly boost the performance of action recognition algorithms and outperform most recently proposed recognition approaches.