StableDiffusion呪文研究：人物の向きとカメラアングル

The script describes a process of using AI for table diffusion with a focus on adjusting character orientation and camera angles through prompts. It mentions using a generative model called Childlike Generative on Google Colab, which is adept at drawing Asians. Additional learning models like Japanese Likeness are loaded to tune the art style. The script instructs on creating an execution environment using a notebook provided by Nolan Ar, which can be accessed through Google Colab. The user is guided to download a zip file, upload it to Google Colab, and select the GPU runtime for processing power. The environment setup begins with the execution of cells in the notebook, and once the URL appears, the user is directed to access the Stable Diffusion Wave UI. The UI allows for image generation by entering a prompt and specifying negative prompts to avoid certain outcomes. The script also details how to adjust the image aspect ratio and use the XYZ plot feature to generate images with up to three parameters. An example is given where the character's orientation is controlled by changing the prompt from 'from front' to 'from side', 'profile', and 'from behind back'.

This paragraph continues the discussion on AI image generation, focusing on adjusting camera angles. It explores different camera angles by changing the prompt to include 'from arrow position' and 'low angle view', and 'from above the birds' eye view'. The script notes that while 'low angle view' did not significantly change the camera angle, 'from above the birds' eye view' successfully produced an image from a top-down perspective. This indicates that certain prompts can effectively adjust the camera angle in image generation.