【初心者🔰】Stable Diffusionのinpaintで『髪型』『髪色』『服装』『表情』『背景』などを修正する方法（コツ）プロンプトや設定を解説

The speaker, Ahiru, returns after a long time and introduces the theme of the video, which is to challenge AI image editing of varying difficulty levels. The assistant, Pegi-kun, is asked to present problems, and they decide to start with a task similar to a first-level magic user exam. The video then delves into the basics of 'painting' in AI image editing, explaining the concept of image-to-image transformation and the potential to create almost perfect images by combining this feature with control nets and image editing software. Two main correction patterns in AI painting are introduced: 'whole picture' and 'only mask,' each with its own use cases and characteristics.

The video provides practical tips on how to use AI image painting tools effectively. It covers the use of different brushes, the importance of painting in one continuous stroke, and the use of rotation arrows to undo mistakes. The segment also explains how to change brush colors, use the eraser tool, and the significance of prompts in guiding the AI's image creation. Settings such as resize mode, mask blur, mask mode, and mask content are discussed, along with their optimal values for different editing scenarios. The video emphasizes the importance of starting with a lower denoising strength and adjusting it according to the desired level of image transformation.

The speaker takes on the challenge of transforming a character's outfit from a maid uniform to a nurse uniform using only AI image painting techniques. They discuss the use of prompts such as 'nurse uniform' and the adjustment of denoising strength to achieve the desired transformation. The video demonstrates the process of iteratively applying paint, generating images, and making fine-tuning adjustments to achieve a convincing nurse outfit. It also touches on the use of the 'fill' tool to help with color adjustments and the importance of seed values in maintaining consistency with the original image.

In this segment, the focus is on adding a smile and a wink to a character's face using AI image editing. The video explains the use of specific prompts like 'smile' and the adjustment of denoising strength to achieve a natural-looking smile. It also covers the challenges of adding a wink to the character's face and how to overcome them by using different painting techniques and seed values. The segment demonstrates the iterative process of generating images, making adjustments, and fine-tuning to achieve the desired facial expressions.

The final part of the video involves changing the background of the character while maintaining the original image's integrity. The speaker explains how to use 'Inpaint Not Mask' to reverse the painting and apply denoising strength to blend the new background with the character. The video concludes with a humorous narrative where the character, Pegi-kun, is surprised by the changes, and the speaker reflects on the capabilities of AI image editing when combined with control nets. The video ends with a playful dialogue between the speaker and the character, emphasizing the creative potential of AI in image editing.