yomoyamastudio








Top Page
Today, now? how you feel.
What kind of picturesque landscape will you paint?
What kind of picture do you want to paint with words?
Exploring psychological states from pictures?




Outline

　This content uses an underlying model of “text” and “image” generation AI. The input is an “image” and the output is also presented as an “image”. The following describes the flow of this process. First, an “image” is used as input, and the result of analyzing the content of the input image is presented as text output. Next, the textual information of the image analysis results is taken as input, and the image is presented as output to reflect the contents of the input textual information.

　We are hopeful that the fact that the output “image” is not directly derived from the input “image” will bring some ideas by including “ambiguity” in the process. This somewhat unreasonable approach poses the difficult task of “reading content” from the input “image” to the generative AI. In other words, since we expect random deviations from the original image, since we have to “read the content from the image” and “generate an image from the words of the read content” in two separate processes, it will be interesting to see what kind of output this approach produces. This may be useful when you want to expand on an idea or image from an existing original image (JPEG or PNG format).

　You can input not only your own drawings, but also your favorite pictures, pictures you are interested in, and so on. It may be interesting to compare the result of the AI with the original image and see how it changes. (Please enjoy the differences in how the various original pictures are expressed. ) You may get some hints or something useful by experimenting, so please feel free to try it out. Upload an image (JPEG or PNG format) as input, and the generated AI will analyze the content of the picture or image and derive an image that will expand your ideas.

Procedure

　Click “Select File” at the bottom to open the image selection window, select the image you wish to send, and click the “Open” button to select the image. Then click the “Send” button to send the image a nd the result will be displayed below the Send button. Please feel free to have fun and try different things. Please note that files sent should be no larger than 2 MB. Please note that large files cannot be processed.