Abstract
In this paper we combine several image processing techniques with the depth images captured by a Kinect sensor to successfully recognize the five distinct human postures of sitting, standing, stooping, kneeling, and lying. The proposed recognition procedure first uses background subtraction on the depth image to extract a silhouette contour of a human. Then, a horizontal projection of the silhouette contour is employed to ascertain whether or not the human is kneeling. If the figure is not kneeling, the star skeleton technique is applied to the silhouette contour to obtain its feature points. We can then use the feature points together with the centre of gravity to calculate the feature vectors and depth values of the body. Next, we input the feature vectors and the depth values into a pre-trained LVQ (learning vector quantization) neural network; the outputs of this will determine the postures of sitting (or standing), stooping, and lying. Lastly, if an output indicates sitting or standing, one further, similar feature identification technique is needed to confirm this output. Based on the results of many experiments, using the proposed method, the rate of successful recognition is higher than 97% in the test data, even though the subjects of the experiments may not have been facing the Kinect sensor and may have had different statures. The proposed method can be called a “hybrid recognition method”, as many techniques are combined in order to achieve a very high recognition rate paired with a very short processing time.
Highlights
In recent years, methods of human posture recognition have been studied in a range of different papers
The remaining notations Ti, i=1, 2,..., 6, indicate the computation time used by each step of the posture recognition process and are defined as follows: T1 is the image processing to remove noise; T2 is the silhouette contour segmentation from the captured images; T3 is the horizontal projection and keeling posture judgment; T4 is the extraction of feature vectors; T5 is the LVQ neural network recognition of the forward-facing sitting, stooping and lying postures; and T6 is the identifi‐ cation of the standing or non-forward-facing sitting postures
It is found that the average success rate of the proposed posture recognition method is higher than 99%
Summary
Methods of human posture recognition have been studied in a range of different papers. Some studies have used a Kinect sensor to recognize human postures; for instance, the authors of [25] presented a method which uses histograms of 3D joint locations from Kinect depth maps and discrete HMM (hidden Markov model) to achieve human posture recognition. Three posture recognition methods involving body width and height ratio, neural network and length ratio are combined to recognize total five postures even when the subjects are facing in different directions. Since it is the fusion of many techniques that helps the method achieve a very high recognition rate in a very short processing time, the proposed method can be called a “hybrid recognition method”.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
