Childhood development students may greatly benefit from the creation and implementation of a cognitive smart lesson that utilizes computer vision technologies. By analyzing and interpreting visual information gathered during lessons in real time, this equipment may be utilized to improve instruction while giving pupils more opportunities for individualized education. Participants' existence, gestures, and movements may be detected in interactive classrooms by installing many sensors and analyzing the resulting pictures via artificial intelligence techniques. This information is then utilized to give pupils more specific input, modify the learning method, and pinpoint problematic areas for improvement. Implementation of such devices in school settings has the potential to raise educational standards by promoting greater pupil involvement and engaged study. In addition to tracking learner growth, instructors may utilize such equipment to pinpoint areas of struggle so they can better customize the material for every learner. One objective of such a study is to create an engaging picture-based study and teaching platform tailored to preschoolers by using intelligent sensing' picture-of-image technologies. Utilizing the algorithm for BP neural networks and the ImageNet, Microsoft Common Objects in Context (MS COCO) , Modified National Institute of Standard and Technology (MNIST), and Chars74K datasets, this article creates a photograph recognition framework that can carry out the majority of its identification acts and integrates it into the educational program using an imagine for pictures. The picture machine learning algorithm has a high precision rate, with a total precision rate of 85.16%. The system offers an improved identification rate, greater instructional effectiveness, and more interactive features than the conventional approach to early childhood education. It has an excellent dynamic impact on learning because it can identify nearly every one of the items kids touch. Early childhood children may benefit greatly from an extra dynamic and individualized educational setting made possible by the introduction of technological vision technologies into a smart teaching. The adoption of computerized visual identification technologies within intelligence classrooms may not only improve the educational and learning experience but also provide vital data for research. Researchers and instructors may learn more about students' learning styles and most beneficial methods of instruction via real-time monitoring of pupil conduct and educational patterns. This information may be utilized to better improve the field of pediatrics by informing the creation of new methods of instruction, resources for learning, and technological tools.