A unified approach to the recognition of complex actions from sequences of zone-crossings

We present a method for the recognition of complex actions. Our method combines automatic learning of simple actions and manual definition of complex actions in a single grammar. Contrary to the general trend in complex action recognition that consists in dividing recognition into two stages, our method performs recognition of simple and complex actions in a unified way. This is performed by encoding simple action HMMs within the stochastic grammar that models complex actions. This unified approach enables a more effective influence of the higher activity layers into the recognition of simple actions which leads to a substantial improvement in the classification of complex actions. We consider the recognition of complex actions based on person transits between areas in the scene. As input, our method receives crossings of tracks along a set of zones which are derived using unsupervised learning of the movement patterns of the objects in the scene. We evaluate our method on a large dataset showing normal, suspicious and threat behaviour on a parking lot. Experiments show an improvement of ~ 30% in the recognition of both high-level scenarios and their composing simple actions with respect to a two-stage approach. Experiments with synthetic noise simulating the most common tracking failures show that our method only experiences a limited decrease in performance when moderate amounts of noise are added.

Mid- to late Holocene Indian Ocean monsoon variability recorded in four speleothems from Socotra Island, Yemen

Four stalagmites covering the last 7.0 ka were sampled on Socotra, an island in the northern Indian Ocean to investigate the evolution of the northeast Indian Ocean Monsoon (IOM) since the mid Holocene. On Socotra, rain is delivered at the start of the southwest IOM in May–June and at the start of the northeast IOM from September to December. The Haggeher Mountains act as a barrier forcing precipitation brought by the northeast winds to fall preferentially on the eastern side of the island, where the studied caves are located. δ18O and δ13C and Mg/Ca and Sr/Ca signals in the stalagmites reflect precipitation amounts brought by the northeast winds. For stalagmite STM6, this amount effect is amplified by kinetic effects during calcite deposition. Combined interpretation of the stalagmites' signals suggest a weakening of the northeast precipitation between 6.0 and 3.8 ka. After 3.8 ka precipitation intensities remain constant with two superimposed drier periods, between 0 and 0.6 ka and from 2.2 to 3.8 ka. No link can be established with Greenland ice cores and with the summer IOM variability. In contrast to the stable northeast rainy season suggested by the records in this study, speleothem records from western Socotra indicate a wettening of the southwest rainy season on Socotra after 4.4 ka. The local wettening of western Socotra could relate to a more southerly path (more over the Indian Ocean) taken by the southwest winds. Stalagmite STM5, sampled at the fringe between both rain areas displays intermediate δ18O values. After 6.2 ka, similar precipitation changes are seen between eastern Socotra and northern Oman indicating that both regions are affected similarly by the monsoon. Different palaeoclimatologic records from the Arabian Peninsula currently located outside the ITCZ migration pathway display an abrupt drying around 6 ka due to their disconnection from the southwest rain influence. Records that are nowadays still receiving rain by the southwest winds, suggest a more gradual drying reflecting the weakening of the southwest monsoon.

Urban Growth Boundaries and their impact on land prices

Undeveloped land transactions at the urban fringe of the Melbourne metropolitan area in Australia are recorded in a dataset that enables exploration of the impact of its urban growth boundary (UGB) on residential land prices. Estimation can take account of a wide range of factors, while controlling for policy anticipation effects and other potential influences on land prices. Modelling estimates indicate that land prices rose substantially inside the UGB after its enactment in 2003, but did not rise much outside of it. These results suggest that the urban growth boundary has had a significant upward effect on the trajectory of the urban region’s house prices. Keywords: urban growth boundary, land prices, land market dynamics

PETS 2015: datasets and challenge

This paper presents the two datasets (ARENA and P5) and the challenge that form a part of the PETS 2015 workshop. The datasets consist of scenarios recorded by us- ing multiple visual and thermal sensors. The scenarios in ARENA dataset involve different staged activities around a parked vehicle in a parking lot in UK and those in P5 dataset involve different staged activities around the perimeter of a nuclear power plant in Sweden. The scenarios of each dataset are grouped into ‘Normal’, ‘Warning’ and ‘Alarm’ categories. The Challenge specifically includes tasks that account for different steps in a video understanding system: Low-Level Video Analysis (object detection and tracking), Mid-Level Video Analysis (‘atomic’ event detection) and High-Level Video Analysis (‘complex’ event detection). The evaluation methodology used for the Challenge includes well-established measures.

PETS 2014: dataset and challenge

This paper describes the dataset and vision challenges that form part of the PETS 2014 workshop. The datasets are multisensor sequences containing different activities around a parked vehicle in a parking lot. The dataset scenarios were filmed from multiple cameras mounted on the vehicle itself and involve multiple actors. In PETS2014 workshop, 22 acted scenarios are provided of abnormal behaviour around the parked vehicle. The aim in PETS 2014 is to provide a standard benchmark that indicates how detection, tracking, abnormality and behaviour analysis systems perform against a common database. The dataset specifically addresses several vision challenges corresponding to different steps in a video understanding system: Low-Level Video Analysis (object detection and tracking), Mid-Level Video Analysis (‘simple’ event detection: the behaviour recognition of a single actor) and High-Level Video Analysis (‘complex’ event detection: the behaviour and interaction recognition of several actors).