I saw the system message the other day about the new image-to-3D tool, but didn’t get a chance to test it out until today. I remember trying the previous version of MakerLab’s Image to 3D tool and wasn’t impressed at all with the results, but understanding that it’s only going to get better, and the latest version has done just that.
While it’s far from being perfect, the output is a huge improvement.
I used two AI generated images and the third is an actual photo, my avatar, which is a picture of me wearing a hoodie that zips all the way up.
The better part of 10 days and 5K filament changes is what I expected for 15 colours. That was sliced using a 0.2mm nozzle to maximize print quality. Completely unrealistic print, but that’s because I’m trying to print this complex model on an A1.
I can’t wait for 32+ colour FDM printers. This tech should be perfected by then.
It’s a great program. I’m doing heads of my family, got four done, tried the fifth and got a “you have insufficient points…”. Site says I have 38 points. Need ten. I must have awful math skills.
I think this tool could be pretty useful in a few ways. I certainly wouldn’t use some of these meshes straight, but they could serve well as a base for figuring out proportions and stuff. That could be pretty useful, especially if I was only making a toy, where proportions/scale doesn’t have to be as spot-on.
Was curious what would happen if I input something from my work, haha. Also something more mechanical like a big dock crane.
It’s shiny because you’re just looking at it in the image-to-3D viewer. It’s not going to print out shiny, and It does still take work to paint the model after it’s been generated.
I’ve watched Image-to-3D and my model be hated on by reddit (some silly stuff) in the past couple days, but no one can tell me that it doesn’t produce good results. I’m literally showing examples of a good result. The detail on it is really good for such a small figure.
Also, I have to wonder what the reaction would have been like if I had posted it with a title like “Just finished making my first model with Blender, how did it turn out?” along with the collage from my last post. I’m willing to bet that it would have been overwhelmingly positive.
I think it’s understandable, people’s reactions. They feel threatened. It feels like you did in a couple of clicks what people have struggled to try and learn for years.
I think all things considered, everything I’ve fed into it has given impressive results. Some of it I could use straight in some manner, some I think would prove to be a great base towards building the model I want
I think it’s a great tool and I am seeing places I can use it within my pipeline.
So now that we are charged just to generate, what do we do when it creates something that wont work? Do we lose our credits? If so, I dont see this feature lasting long, I refuse to use my credits with only hope that it’ll look the way I want. Did a cabinet, missing the back feet so now either I pay again, hope that the next model fixes this, or look for another replacement. Very sad, I understand why they need to charge, but there should be a way to protect those of us after we pay to redo the print again, etc.
