Human pose classification within the context of near-IR imagery tracking

Abstract

We address the challenge of human behaviour analysis within automated image understanding. Whilst prior work concentrates on this task within visible-band (EO) imagery, by contrast we target basic human pose classification in thermal-band (infrared, IR) imagery. By leveraging the key advantages of limb localization this imagery offers we target two distinct human pose classification problems of varying complexity: 1) identifying passive or active individuals within the scene and 2) the identification of individuals potentially carrying weapons. Both approaches use a discrete set of features capturing body pose characteristics from which a range of machine learning techniques are then employed for final classification. Significant success is shown on these challenging tasks over a wide range of environmental conditions within the wider context of automated human target tracking in thermal-band (IR) imagery.

Posture estimation for improved photogrammetric localization of pedestrians in monocular infrared imagery

“…Future work will look to investigate the extension of this approach to the recovery of multiple pose attributes [14,40], as an enabler to human activity classification [6,41,42] and also into visible-band imagery using recent advances in real-time salient object detection [43]. Applicability within the context of mobile platform navigation [44][45][46][47], driver assistance systems [48][49][50] and for multi-platform, multi-modal wide-area search and surveillance tasks [7,51,52] will be further explored.…”

“…The work presented in this paper is a direct extension of [1] that demonstrates photogrammetric pedestrian localization within thermal-band imagery incorporating a lightweight tracking solution akin to that of [6]. In [1] photogrammetric pedestrian target localization is presented to an accuracy significantly within the commonly regarded "gold-standard" of consumerlevel Global Position System (GPS) positioning (typically ±5m under ideal conditions [29]).…”

“…The posture of a pedestrian target detected within the scene is estimated as a percentage of full height (maximal posture) based on the use of a Histogram of Orientated Gradient (HOG) feature descriptor extracted from each detected scene target and the use of Support Vector Machine (SVM) based machine learning regression. In contrast to prior work in the field, we leverage the key advantages of thermal-band infra-red (IR) imagery for pedestrian localization with a tracking framework [6] and demonstrate robust target localization, independent of variations in target posture (i.e. behaviour), within the statistical error bounds outlined in [1].…”

“…A key advantage of such thermal-band infrared (IR) imagery for pedestrian localization is both androbust detection of human shape signatures within the scene [5][6][7] and robust localization of their scene bounds in pixel-space (e.g. Figure 1).…”

“…As such, the principles of photogrammetry can be used to recover 3D pedestrian position within the scene based on a known camera projection model and an assumption that variance in human height is in fact quite small (statistically supported by [8,9]). In [1] we experimentally investigated the accuracy of classical photogrammetry, within the context of current target detection and classification techniques [5][6][7], as a means of recovering the true 3D position of pedestrian targets within the scene. A real-time approach for the detection, classification and localization of pedestrian targets via thermal-band (infrared) sensing was presented with supporting statistical evidence underpinning the key photogrammetric assumptions.…”