• Aashna Mahjan
• Shujath Khan
• Arjun Ramesan

The spinal cord of the human body serves as the integral part to sustaining our body’s posture. An incorrect posture may lead to back and spinal injuries. To avoid this, we have come up with an application that uses Computer vision and Neural networks to analyze the human postures and immediately notify the users to recheck their postures through an Android app and a Desktop app. The data provided as input to the network is divided into two sub-classes and each posture is then recognized and analyzed. For primary training and testing purposes we have used Convolutional Neural Networks as it provides this intuitive way of looking at images, analyzing adjacent pixels in small areas and input those readings in a pooling layer. For final training and testing purposes we have used GANs, Generative Adversarial Networks, as it uses a generator network for generating the correct and incorrect data and dynamically classifying them. We tested the entire working model using a real-time video feed of various environments, the model proves to be highly accurate.

Posture Recognition; Neural Network; Feature Extraction; Image Processing; Python; Computer Vision; CNN; GANs; Android Application.

"Real-Time Human Posture Recognition", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.5, Issue 11, page no.445-452, November-2018, Available :http://www.jetir.org/papers/JETIR1811569.pdf

"Real-Time Human Posture Recognition", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.5, Issue 11, page no. pp445-452, November-2018, Available at : http://www.jetir.org/papers/JETIR1811569.pdf