• Sri Ram Polisetty
• Saikumar Itraju
• Harshini VSS
• Pavan Kumar K
• Hari Krishna V

In the present scenario human requires security services like hospital, police and fire services. These services mainly depend on the human intervention. Without human interaction these services will not respond automatically. Due to lack of quick human interaction, cost of damage increases. In order to reduce the damage, we co-ordinate these services through Artificial Intelligence. Based on the intensity of the damage, photographs and exact location is sent to the respective departments. In this work, we investigate the automatic detection of fire pixel, accidents, and Illegal threat regions in video imagery within real-time bounds without reliance on temporal scene information. As an addition to previous work in the arena, we deliberate the performance of experimentally defined, reduced complexity deep convolutional neural network (CNN) architectures for this task. Opposed to modern trends in the field, our work demonstrates utmost accuracy for entire image binary fire detection, with maximum accuracy within our super pixel localization framework can be attained, via a network architecture of considerably reduced complexity. We show the relative performance accomplished in contradiction of prior work using standard datasets to demonstrate maximally strong real-time fire pixel, accidents, and Illegal threat region detection.

Artificial Intelligence, pixel, convolutional neural network, super pixel localization framework, datasets

"AUTOMATIC DETECTION OF FIRE AND ACCIDENT PIXELS USING TENSORFLOW TECHNOLOGIES", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 4, page no.134-142, April-2019, Available :http://www.jetir.org/papers/JETIRBE06026.pdf

"AUTOMATIC DETECTION OF FIRE AND ACCIDENT PIXELS USING TENSORFLOW TECHNOLOGIES", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 4, page no. pp134-142, April-2019, Available at : http://www.jetir.org/papers/JETIRBE06026.pdf