Transforming Digital Images Into Cartoon Styles Visual Using Convolutional Neural Networks

The art of cartoonization in images has garnered significant attention, especially with the breakthroughs witnessed in entertainment, digital art, and social media.The traditional methods rely on the carted filters and edge detection techniques, which sometimes fail to retain the important features in the image.This paper proposes a deep learning-based technique using convolutional neural network (CNN) to apply this cartoon effect.Readjusted description of CNN addition works-detection, feature extraction from real-world images, edge enhancement, color quantization, and finally achieve quality outputs as outputs.A diverse dataset of real as well as cartoon images was used for training our model, while it was optimized by loss functions such as MSE and perceptual losses. The proposed approach is evaluated for structural similarity, perceptual quality, and computational efficiency. Indexed databases demonstrated that our approach is better than traditional approaches as it creates appealing cartoon-type outputs while significantly keeping content integrity intact. This work will contribute to an ever-growing time with great relevance to various AI-enabled artistic transformations which can be of great use in digital media, animated storytelling, and creative content generation.