Asking Gemini to evaluate a photo with regards to it being featured in Google Maps


@AdamGT has been working on a three part post where guides are asked to choose which image of the two or 3 presented is most likely to get the most views on a Google Maps POI.
In the discussion of the third part I raised the idea of asking what Gemini would have to say about this. I asked Gemini two questions

  1. how does Google maps decide which photo to feature in a business’s gallery
  2. how does a computer evaluate the elements of a photo
    The second question is more relevant to Adam’s post.
    @AT_Rome did something much more sophisticated: he created a prompt asking Gemini to act as a “analyst for Alphabet’s Google Maps division.” and described the 3 photos for it to consider. Here are his two posts about it.
    First response

Details of prompt description
When I wrote up about my Gemini questions, I added a third question:
How likely are my photos likely to be featured in Google Maps?
I did not actually ask the question but made up a fictitious response. To try and prevent anyone from thinking the reply was genuine, I translated the text into Klingon (a world & language from the TV show Star Trek) so that most people wouldn’t be able to understand it.
But @AT_Rome 's thoughtful work made me think: “Why not actually ask Gemini to look at a photo?”
So, I submitted the photo shown above with this request:
evaluate this photo to see if it would be featured in the business Google maps gallery. the maps link is Panda North
Below is the response that I received:
Evaluation of the Photo for Google Maps Business Gallery
Here’s an assessment of the photo, keeping in mind Google Maps’ focus on showcasing businesses:

Strengths:

** Appetizing Subject: The photo features a well-presented food dish, which is highly relevant for a restaurant’s Google Maps listing. Food photos are often a key factor in attracting customers.*
** Good Composition: The dish is the clear focus, and the arrangement with the rice, kale, and carrot garnish adds visual interest.*
** Decent Lighting: The food is well-lit, showing its texture and colors nicely.*
** Clear and Sharp: The image appears to be in focus, allowing viewers to see the details of the dish.*

Potential Improvements:

** Angle: While the current angle is good, a slightly lower angle might make the dish appear even more prominent and appealing.*
** Background: The background is a bit blurry, but it’s not distracting. However, a cleaner, less cluttered background could further enhance the focus on the food.*
** White Balance: The white balance seems slightly warm. Adjusting it to be more neutral could make the colors of the food look even more vibrant and accurate.*

Likelihood of Being Featured:

** The photo has a good chance of being accepted and displayed on a Google Maps business listing. It meets the basic criteria of being relevant, visually appealing, and informative.*
** To increase its chances of being featured in a prominent position (or even in highlights), addressing the potential improvements mentioned above would be beneficial.*

Overall:

The photo is a positive representation of the food and could be a valuable asset to a restaurant’s Google Maps presence. With some minor adjustments, it could be even more effective in attracting customers.
In the two days since I added the photo to Maps it has gotten 237 views.
I think the analysis is reasonably accurate and it appears that one could use Gemini as a way to improve one’s photography for Google Maps.
It would be an interesting project for a guide to take photos at multiple POI’s, submit them for review by Gemini and then go back and retake the photos based on Gemini’s suggestions for improvement. And then submit the original and the second photo to see how they fare against each other,

29 Likes

Brilliant technique @Rednewt74 , lovely, I am gonna do that right away will ask Gemini to check few of my pictures and see what improvements needed in the photos I click

1 Like

Nice to see after very long time dear @Rednewt74 & your Analysis always super :folded_hands:

3 Likes

That is good techniq @Rednewt74.
Thank you for sharing information about the experiment with Gemini

1 Like

Interessante Analyse und Beschreibung @Rednewt74

It’s a great technique @Rednewt74 I’m surely gonna try it. Thank you for sharing

1 Like

Firstly, it’s a nice photo with well presented food on the plate @Rednewt74. As you know based on the interesting approach that @AT_Rome took in researching his reply for the A, B or C - Which photo has the most views #03 post, in which he utilized the experience he gained with prompt engineering, I thought that he would take this as the next step but well done you on taking the initiative.

In my view, while this could be a useful approach for generating potential ideas on how to improve one’s photos (e.g. the potential improvements that you mention in the case of your plate of food), personally I that this would be a bit after the horse has bolted and don’t think it’s practical for Local Guides to go to so much trouble gaining knowledge on how they might be able to improve on a photo already taken (e.g. download it, edit it and re-upload) or take another photo based on the suggested improvements.

However, their can be little doubt that such AI could be incorporated in the image editing tools provided in say Google Photos, the camera app or even Google Maps itself so that prior to uploading onto Maps, image improvements could be made. For example, in the case of your photo the AI editing could handle the “blurry” background and the “white balance” issues just prior to you uploading it.

1 Like