Sensor Based Human Gait Recognition for Drunk State

Abstract

Many research studies have examined human activity recognition and body posture and gait detection. Relatively fewer studies have investigated specifically into shaky, abnormal gait or how drunken person move in a low lit, uncontrolled setting. This research paper uses markers for tracking the body joints and legs, hands and face and used the position and movement of tracked features to train an SVM classifier for recognizing various abnormal actions while walking such as tripping, falling, shaking, walking side-ways, falling, dragging and walking with helps from others vs normal gait and standing posture. 5 subjects participated in script based simulated drunk actions, expressions and behaviour from a list of actions that represented an abnormal gait and drunk state of mind. The classification results showed 72.8% accuracy under controlled lighting. The accuracy was lower by 3.1% in dim indoor lighting and natural outdoor setup similar to those outside a bar or a low lit street at night.

Authors and Affiliations

Amol S Patwardhan

Keywords

Related Articles

Effect of Parasitic Inductance on Inverter Module Performance

Effect of circuit parasitic inductance is an important consideration in application and characterization of Insulated Gate Bipolar Transistor Module (IGBT). Higher value of parasitic inductance causes higher voltage ove...

Section Optimization of Diaphragm Wall.

Diaphragm wall is common type of earth retention system. Diaphragm walls are generally used in deep basement of building, congested urban spaces, riverfront structures and marine structures. The stability is provided th...

High Density of Noise Removal Using FBD Filter

This research work has been done on data mining for noise type detection and filtering process to ample more effectiveness for removing the noise content in an image. In this article we have proposed a scheme of data mi...

Motion-Based Detection and Tracking of Moving Objects in Real Time Scenario

Detecting and tracking moving objects in real time scenario is a vast, active and complex area of computer vision. Detection and tracking is used in many fields like surveillance, intelligent tracking system and many ot...

Implementing user-defined Integrity Constraint in MYSQL

The methods with which we can maintain correctness of any specific table in data base management system are Integrity constraints. The general idea of this paper is to investigate and observe user-defined integrity cons...

Download PDF file
  • EP ID EP22627
  • DOI -
  • Views 202
  • Downloads 4

How To Cite

Amol S Patwardhan (2016). Sensor Based Human Gait Recognition for Drunk State. International Journal for Research in Applied Science and Engineering Technology (IJRASET), 4(9), -. https://europub.co.uk/articles/-A-22627