Briggs's post
cancel
Showing results for 
Search instead for 
Did you mean: 
Level 10

Identifying the content in an image prompt

Caption: The prompt when uploading photos to identify the image's contents.Caption: The prompt when uploading photos to identify the image's contents.

Just a few days ago, I had a journey along the Great Western Highway to walk about 12km west while snapping some photos for my personal endeavours. It wasn't until about yesterday (Saturday) that I started uploading photos on the Google Maps android application, and it's come to a point where these prompts are starting to become overbearing, perhaps excessive to an extent.

The nature of my photos often don't protrude to actual activities nor to the macrophotography of products or people. Therefore, identifying an activity, product or menu item in most of my uploads is not necessary or accurate as most of the time my photos are physical, landscape and/or environmental.

 

Here are a few examples where there has been a disconnect in these prompts, including the above examples:

  • For a Restaurant/Cafe/Food store, it asks what menu item the photo has. I usually take photos of the shopfront, dining space, counter area and the general ambience of the place.
  • For an Association or Organization/RSL Club/Sports Club, it asks what product the upload has. This is generally irrelevant as these are usually entertainment or social gathering areas with the possibility of a few eateries or such.
  • While it does ask what product a Supermarket, Convenience store or any other retail store the upload wishes to be identified, once again the majority of my photos are either the shopfront, some shelves and the interiors of the store. For most local guides, their photos within these stores will rarely consist of just one product. I don't see the premise of local guides uploading just a can of tuna, or a bunch of bananas, or just a pair of jeans.
  • For any sort of Park/War Memorial/Sports Field or Public Space, it prompts and asks what activity is present in the photo. I can understand if there is a feature such as a rock-climbing wall or a slide by itself, but generally if there are no people in the photo either cycling around or playing sports, there's not much leeway to state the activity. I'm also not sure what sort of activities might also occur in a War Memorial (pictured above).

I can think of an idea or two regarding this feature, as this hasn't necessarily meshed well with my use of Google Maps. While the sole function I used this feature was to identify the dish served in a restaurant and then either to recommend it or not, I don't see myself identifying activities in a park, nor taking photos of singular objects in a supermarket or store. So I would request:

  • The option to turn off these prompts of identifying the item upon upload within the app.
  • The ability to identify and only prompt food when uploading, and distinguish whether a photo of the restaurant/shopfront is being uploaded, instead of having all uploaded photos prompted based on a listing's category alone.
  • The better usage of AutoML (Machine Learning) and labelling of photos during the upload process. There was an article about Google's AutoML identifying the actual restaurant that the dish came from which I thought was fantastic. If you're on Google Photos, this is already pre-done to a great extent just by searching either 'Food' or 'Restaurant' in your collection.

Anyway, that's it for now regarding this feature. While I do enjoy identifying food and recommending such (in contrary to encouraging captions on photos), when it asks me to identify food when I upload a picture of a few tables, I don't think that's relevant. As per the other examples, that's pretty the gist of things.

3 comments
Level 10

Re: Identifying the content in an image prompt

Google's image labelling through it's machine learning and crowdsourced image verification has really allowed the software to evolve to great heights. Even photos that I have only taken a photo of a few seconds ago are already labelled with its features, even to a point where searching 'food' will index the photo immediately in Google Photos.

 

Caption: What image suggestions might look like, as the AI suggests the image contains a Chicken Burger.Caption: What image suggestions might look like, as the AI suggests the image contains a Chicken Burger.

So I recommend allowing Google Maps to suggest the dish/product identified in the photo instead of having to always prompt and select the dish after typing it in. This saves time, it's convenient and it instills usefulness to the user. Of course, once a user starts typing a product name and selects one, the suggestion is removed.

Former Google Contributor

Re: Identifying the content in an image prompt

Hey @Briggs,

 

Thanks for this one. Forwarding along! Your feedback is definitely important. Also, I'm sorry for any inconvenience caused because of the photo prompts in your Maps application. 

Level 10

Re: Identifying the content in an image prompt

Thanks for forwarding the feedback along Arielle.

It's okay, I've managed to start bypassing all the prompts by clicking the 'back' button on my Android phone. It works like a confirmation, but that's pretty much it.