Content Aware Align

[RiskyTrizkit]

Using AI recognition to be able to align certain elements to the canvas or to eachother would be beneficial for my workflow. If I have for example 100 images of people in different poses, and need to align their faces to the center of the canvas, that is a lot of eyeballing and manual work. However, with AI recognition I believe we should be able to do this now.

Faces would be the obvious use case, but if it would be possible to customize this even further that would be spectacular. For example, typing in what to search a batch of images for, such as "XYZ Logo". 

[Tracy.H]

Thank you for the feedback. Could you explain this to me a little more? Are you trying to create 100 separate images and you want them all to have the same alignment? Or are you wanting, say, 100 people in an image and make sure they are all positioned and aligned similarly?

Also I'm not sure I understand your use canse of XYZ logo. We currently don't train on copyrighted materials so we wouldn't recognize a brand name logo. Or are you saying you want to create a logo in the image?

[RiskyTrizkit]

Say I have 100 separate images of people, or 100 layers on one single file. (One person's portrait per layer) The head of each person could be in any general position on the XY coordinates of each image per layer, and at any random size. The objective would be to batch align the head/face of each individual on each layer to the center of the canvas using AI recognition. (Since AI can tell what faces are) 

Or for example, say you have 100 different studio photos of 100 different cars - 100 layers. Perhaps several cars are offcenter in the photograph or sized differently in each layer. Rather than manually resize and manually align to center of canvas, there could be an option to use AI to batch run this process using AI's knowledge of what a car looks like.

This would be an outlier use case but it is mine particularly - I work for a AAA video game that has an in game store. we have over 1000 different tanks that we display. One struggle is aligning each tank so their body size (hull size) matches up in relation to the rest of the tanks. This is an incredibly manual but yet very important task. If we were able to have an ability to train the AI ourselves (or just to learn the to learn what a chassis is, or what a hull is, and what a cannon is, and assist in aligning and resizing, possibly even rotating if needed that would really be next level helpful. 

[Rachael4]

Were you ever able to find a program that does this or figure this out?