Abstract:
GAN is an emerging machine learning technique that has the ability to convert text into an anime sketch. Anime Sketch Generator using Text Description is a hybrid of NLP and computer vision. Combining two data types to complete this task is a challenging task. A recent image captioning project that utilized the multi-model architecture, on the other hand, illustrated that we can accomplish this task. Image captioning is the reverse process of Anime Sketch Generator Using Text Description (Image to Text). Anime Sketch Generator Using Text Description can benefit and be useful to generate an anime sketch for an artist who has a vague idea about a required character and intends to fill in the gaps in the thought process, which can take a long time in real life due to the duration of the thought process. The project’s main goal is to create an android application and train the dataset with GAN to demonstrate suggested applications practically. The app will run on Android smartphones with internet access. All image loading, model saving, photo generation, and user feedback modules are fully functional. However, we are unable to generate a high-quality photograph because the quality of the generated image is decided by the quality of the training dataset.
