So, if you are using the recorder, then there is a selector switch in the top right for image based or UI based recording, but I don't think that's your case.
During "Capture UI element", you get the UI selectors. It also saves an image, not because that is what it is using, but it's for you, the developer, to know what that is supposed to be clicking. The photo doesn't actually have an impact on the bot if you are using selectors.
With that said, I think your best bet will be to "Move Mouse to text on screen (OCR)." Use the date as the "Text to find variable" (make sure it is formatted correctly M/d/yyyy). Under Advanced, use "Send a mouse click after moving mouse" and "mouse position relative to text"...my guess is "Left" Offset X "-15" or so; you can play around with it until it works, but the idea is "find the date and click the box on the left."
Good luck.