MicrostockGroup Sponsors

Envato Elements

Author Topic: AxonGuru - auto tagging and auto categorize with latest AI and cloud technology  (Read 2389 times)

0 Members and 1 Guest are viewing this topic.

axon_guru

  • Save time with Axon Guru!
« on: September 19, 2016, 14:54 »
0
Hi Guys, we have created a new site to streamline your stock photo submissions. It will save you 2-5 minutes per image! Just host your photos on one of the leading photo hosting sites (Google Photos, Amazon CloudDrive, Dropbox, Smugmug etc.), and we will do all the heavy lifting. We will recommend the marketable photos, tag them automatically and pre-fill categories automatically. We even let you compare your photo to similar ones already on the marketplace. All you have to do is review and submit. There is no FTP or sharing passwords. It is using the latest AI and cloud technology. It is free for basic plan with unlimited low priority processing right now.  http://axonguru.com

We are also looking for testers and reviewers. Drop me a line if you are interested.


« Reply #1 on: September 19, 2016, 15:30 »
+1
Looking at your tutorial page - it'd be nice if your examples were larger as you really can't see much from those tiny screenshots - I don't think you can possibly do the sort of work needed to manage stock photos automatically

https://guru.axonimage.com/tutorial

What is "Auto Awesome"? Some variation of the automatic contrast and color balance that apps like Photoshop provide? Fine for selfies and instagram but not if you want to submit stock for sale. I find it hard to accept that there is any way you can sift through images to decide which are good ones to upload - what are your criteria? I'm not asking for trade secrets, just some sort of overview.

With keywords, your example of a tree is pretty straightforward - how do you deal with lifestyle images or those where a specific location or other technical information is essential? Shutterstock's auto-keywording feature is probably helpful for some images (particularly for a non-native speaker) but I can't imagine that you can tell Portland Maine from Portland Oregon or handle a multi-ethnic business meeting or food prep or...

And do you handle releases?

There's only a time saving if the information provided is any good - otherwise you have an automated way to consign images to oblivion. Good keywords are critical.

« Reply #2 on: September 19, 2016, 22:21 »
0
Looking at your tutorial page - it'd be nice if your examples were larger as you really can't see much from those tiny screenshots - I don't think you can possibly do the sort of work needed to manage stock photos automatically

newbielink:https://guru.axonimage.com/tutorial [nonactive]

What is "Auto Awesome"? Some variation of the automatic contrast and color balance that apps like Photoshop provide? Fine for selfies and instagram but not if you want to submit stock for sale. I find it hard to accept that there is any way you can sift through images to decide which are good ones to upload - what are your criteria? I'm not asking for trade secrets, just some sort of overview.

With keywords, your example of a tree is pretty straightforward - how do you deal with lifestyle images or those where a specific location or other technical information is essential? Shutterstock's auto-keywording feature is probably helpful for some images (particularly for a non-native speaker) but I can't imagine that you can tell Portland Maine from Portland Oregon or handle a multi-ethnic business meeting or food prep or...

And do you handle releases?

There's only a time saving if the information provided is any good - otherwise you have an automated way to consign images to oblivion. Good keywords are critical.

Hi Jo,

Thank you for your comment. I'm also one of the creators of Axon Guru. It is true that we can't say too much about the criteria used in selecting user's photos. What we can tell you is the photos we picked up share common features with those enjoying best selling records on StockPhoto websites. This is made possible by deep learning technologies. This criteria alone will not ensure the approval of the photo, but it will increase the chance of approval for new contributors. For experienced contributors, they may have better equipment and skill, but can still benefit from the recommendation process based on sales data and the metadata auto-completion feature. Meanwhile, we are also working on some noise level, lighting / composition detector to improve the chance of approval, which can further improve the submission experience.

We may not be able to pick up unique, never seen photo from user's collection, but we do have an option to let the user pick their favorite photo manually, and still benefit from our deep learning based metadata completion technique, as well as comparing with other similar best selling photos. To our best knowledge, this is the only tool that can automatically fill out all the necessary metadata for multiple popular stock websites.

About the accuracy of metadata: For keywords, in most recommended photos we tested, the automatically generated keywords and categories are usually accurate, and when needed only very few tags need to be deleted (a very easy step). Even when user would like to provide his/her unique ones, it is still much easier than having to come up with dozens of keywords from scratch. The suggested title and description are often very useful examples to help users come up with their own.

We currently do not handle model or property release (but we do remind user if we think a release is needed), but we are happy to make it a high priority if this feature is in high demand. Please feel free to suggest more features with the feedback function in our tool. We would also be happy to learn from you about other alternative tools.

Hope you will enjoy using Axon Guru tool.


 

Related Topics

  Subject / Started by Replies Last post
17 Replies
5554 Views
Last post January 25, 2007, 14:31
by pixelbrat
2 Replies
1831 Views
Last post June 15, 2012, 08:01
by hjalmeida
3 Replies
1892 Views
Last post December 28, 2012, 23:48
by RacePhoto
3 Replies
1219 Views
Last post June 04, 2013, 21:11
by klsbear
6 Replies
3367 Views
Last post April 03, 2015, 01:36
by fmarsicano

Sponsors

Microstock Poll Results