In our daily lives, object identification has grown in importance as a task for a variety of objectives. AI techniques have been used in the past for this task, but they have currently been used to the arranging of species based on pictures in order to extract the capabilities list. Selecting the appropriate article discovery is aided by this list of capabilities endeavor. This application suggests a cooperative learning-based profound learning method to overcome the article order problem. In this study, the unique convolutional neural networks (CNN) are focused. Below, the voting portion of the voting plot is used to better the outcome. General work on the CUB 200-2011 dataset is finished. When compared to the unique CNN models, the results revealed an astounding increase in the accuracy of the suggested work. Here, the evaluation makes use of Japanese humor, or manga.