Smoke-Free Zones: Enhancing Smoker Detection with MobileNetV2 and CNN

Smoking is still a major global public health concern since it causes a lot of avoidable illnesses and early deaths. Planning an intervention and a course of treatment effectively depends on accurately identifying smoking behaviour. With the help of MobileNetV2, a portable and effective deep learning model, the convolutional neural network (CNN) architecture in this work presents a unique method for smoker categorization. This method minimizes computational resources while improving classification performance by utilising the hierarchical features that MobileNetV2 learned. Model refine the pre-trained MobileNetV2 model on the smoker categorization task by transfer learning. The efficacy of suggested method is demonstrated by the experimental results, which achieve cutting-edge performance in terms of smoker classification precision, sensitivity, and specificity. Moreover, model demonstrates resilience in many data modalities, indicating its potential for practical use in healthcare environments. A viable method for automated smoker classification is provided by the suggested CNN + MobileNetV2 framework, which will help with early diagnosis, personalized intervention, and better patient results in attempts to quit smoking which could lead to healthy lives.