Gait recognition in the wild using shadow silhouettes

Abstract

Gait recognition systems allow identification of users relying on features acquired from their body movement while walking. This paper discusses the main factors affecting the gait features that can be acquired from a 2D video sequence, proposing a taxonomy to classify them across four dimensions. It also explores the possibility of obtaining users' gait features from the shadow silhouettes by proposing a novel gait recognition system. The system includes novel methods for: (i) shadow segmentation, (ii) walking direction identification, and (iii) shadow silhouette rectification. The shadow segmentation is performed by fitting a line through the feet positions of the user obtained from the gait texture image (GTI). The direction of the fitted line is then used to identify the walking direction of the user. Finally, the shadow silhouettes thus obtained are rectified to compensate for the distortions and deformations resulting from the acquisition setup, using the proposed four-point correspondence method. The paper additionally presents a new database, consisting of 21 users moving along two walking directions, to test the proposed gait recognition system. Results show that the performance of the proposed system is equivalent to that of the state-of-the-art in a constrained setting, but performing equivalently well in the wild, where most state-of-the-art methods fail. The results also highlight the advantages of using rectified shadow silhouettes over body silhouettes under certain conditions.

Automatic Classification of Gait Impairments Using a Markerless 2D Video-Based System

“…The normalization step, applied to each frame, makes the proposed system robust to scale changes, such as those resulting from a varying distance between an individual and the camera. The distance between the feet of the individual is then approximated as the width of a rectangular bounding box fitted onto the silhouettes [ 33 ]. Using the silhouettes and the distance between the feet, the proposed system can compute the desired biomechanical gait features.…”

“…The proposed system selects the torso using the human anatomy ratios presented in [ 34 ]. It then performs principal component analysis over the torso and measures the angle ( θ ) between the horizontal axis and the first principal component, ( , ), according to Equation (10) [ 33 ]: …”

Deep Gait Recognition: A Survey

“…In order to help illustrate an overall structure for the available gait recognition approaches, a few taxonomies have been pro- posed in the literature [18], [38], [64], which have organized the available solutions from different perspectives. The taxonomy proposed in [38] is based on the type of sensors, classifiers, and covariate factors such as occlusion types.…”

“…The taxonomy in [64] categorizes gait recognition methods based on the type of features used. Finally, the one proposed in [18] considers user appearance, camera, light source, and environment-related factors.…”

“…The performance of vision-based gait recognition systems, hereafter only referred to only as gait recognition, can be affected by i) variations in the appearance of the individual, such as carrying a handbag/backpack or wearing items of clothing such as a hat or a coat; ii) variations in the camera viewpoint; iii) occlusion factors, for instance where parts of the subject's body are partially covered by an object, or alternatively by a part of the subject's own body in certain viewpoints (also known as selfocclusion) [15], [16]; and iv) variations in the environment, such as complex backgrounds [17] and high or low levels of lighting [18], which generally make the segmentation and recognition processes more difficult.…”