Early fire detection technique for human being using deep learning algorithm

Fire and smoke detection in today’s world is a must, especially clustered areas where quick response can prevent significant damages save lives. Early plays role preventing the fire from spreading by alerting emergency personnel. It may not be possible to install traditional detectors everywhere. As result, incorporating into existing closed circuit television (CCTV) systems various places provide warning appropriate authorities, allowing for action spreading. This work aims developing an early prediction model with CCTV footage images video frames. The videos are collected multiple datasets available online. A convolutional neural network (CNN) developed prevention of compares it transfer learning models ResNet50 VGG19. obtain accuracy around 94% using CNN model, 95% VGG19 98% ResNet 50. high replace which both cost-effective easy implement surveillance cameras.