Surveillance cameras play a crucial role in upholding public safety, safeguarding properties, and acting as a deterrent to criminal activities. However, the conventional surveillance systems heavily rely on basic image processing techniques and rule -based algorithms for tasks like object detection, tracking, and recognition. Unfortunately, these methods have their limitations when dealing with complex scenarios such as occlusions, changes in lighting conditions, and variations in object appearances. Therefore, false alarms or missed detections can occur, ultimately diminishing the overall effectiveness of the surveillance system. To tackle these challenges head -on, researchers and engineers have been actively exploring the integration of advanced computer vision techniques, particularly Convolutional Neural Networks (CNNs), to augment the capabilities of smart surveillance cameras. The aim is to improve the efficiency and accuracy of these systems by leveraging cutting-edge CNN modifications. The need for enh ancing the performance of smart surveillance cameras arises from the ever -growing demand for more intelligent and reliable surveillance solutions. As the number of surveillance cameras deployed in public spaces, commercial areas, and private premises conti nues to increase, it becomes imperative to develop advanced algorithms capable of accurately and efficiently analyzing vast amounts of video data. By elevating the performance of surveillance cameras, it becomes feasible to provide real-time and accurate i nsights to security personnel, law enforcement agencies, and other stakeholders. Hence, the core objective of this project revolves around enhancing the performance of smart surveillance cameras using advanced CNN modifications. A custom CNN architecture i s proposed to optimize object detection, tracking, and recognition tasks. This approach undergoes meticulous experimentation with real -world surveillance data to evaluate the effectiveness of the modifications. The results manifest significant improvements in the system's accuracy and efficiency, paving the way for more intelligent and reliable surveillance solutions across various applications. The outcomes of this research hold the potential to contribute significantly to the fields of computer vision and smart surveillance technology, enhancing public safety and security overall.