3D Scanner: App to turn 2 photos into 3D model

Dear ALU! Thank you for sharing your experience.

I just used this program today and was very pleased with its reliability: 8 out of 8 scans turned out good.

I am currently preparing a video "Scanning Christmas toys". I bought 7 things in the supermarket (stuffed toys, gingerbread, porcelain figurine).

I took two photos of eight toys in 16 minutes and got eight 3D models in 19 minutes with this program.

I have processed with default settings for now, so the first models turned out quite rough - only 100x100 pixels. But I see that the models are good:
https://ibb.co/XyXn2xB

Tomorrow I will reprocess them in 400x400 resolution, and apply a texture. Then the models will be very beautiful, like here:
https://ibb.co/Tcjzkxy
And I will print them out and make a video.

Here are my pairs of photos from today, which anyone can process and get good models.
https://drive.google.com/file/d/1_Z1kxcyJcDtjtX4YRXolfBJ5AKxGTBjj/view?usp=sharing

Some people complain that they get peaks or holes in the models. I don't really understand these complaints. Mr. Diester recently wrote here that I am a magician, and only I can make models. I think there is only one secret to magic - textures.
1) We get good scans if we scan an object with a good texture on a background with texture.
2) We get good scans if the object or background was textureless, but pre-textured (painted, pasted motley papers, sprinkled with something)
3) We get bad scans if the object or background has no obvious texture (shiny, single-color, translucent)

I am not saying that this scanning method is ideal. More powerful scanners are needed for engineering applications. However, for decorative purposes, my program produces excellent scans. I believe that this simple software scanner should be in the arsenal of any 3D printing person.

And I am very pleased that today I received 8 scans out of 8 that were quite excellent to make a good video)

Thank you for your suggestion regarding examples and presets! Now there is this problem: by default, the gets a rough 100x100 model without texture, and the does not realize that by setting 400x400 and a 3 mm texture, one can get very beautiful models. I will fix this problem with presets in the next program update.

Moreover, in early 2025, I plan to release version 2 of this scanner, which will have four powerful improvements. First of all, this will be aimed at obtaining excellent scans from films and computer games. This will also simplify the scanning of ordinary objects. Once you buy the program, you will be able to the improved version for free.

Thank you for your !

Dear Diester!
Thank you for your experiments with this program

I have already released 17 programs, but this program requires the to understand the principle of operation and be careful. The program requires the to take two photos with a shift.

However, by ing two photos that do not match each other, it is impossible to get a 3D model, no matter how perfect the program is, and no matter how much I refine the program. Therefore, the refinement of the program was suspended, because the issue is not in the refinement. The problem is how to help the take two photos that will produce a good 3D model.

This is a good program and you can get scans no worse than I show in photos and videos. But to do this, you need to do what I did. For this, I specially made illustrated instructions in PDF. And besides, I am always happy to explain in detail how to achieve good quality scans.

no but you really have to stop taking people for idiots. read what people write. Many complain about it. And even with the photos given on google drive it doesn't work. When I asked for a refund, I was promised an update last year that was supposed to fix the problems, but it was never fixed and the excuse is that you're a beginner... Personally, I've given up and I use Meshroom, which works perfectly and is free.

What shocks me the most is that people are still being duped. And that this programme is still here.

So that you do not slander publicly. Please process these 4 sets of photos using the program and post the results here.

https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link

If you see that the results are good - then the problem is not in the program (the program is excellent), but in the fact that we need carefully taken photos, with an understanding of the stereo principle.

You are not happy with my program (no other technology on this planet will combine 2 photos better than my program).

You are unhappy with the fact that you need to take photos with a accurate shift, with uniform lighting, with preliminary texturing of bad areas (mirrored, transparent or monochrome) of the object that give peaks.

As I said, I've given up on this programme, which doesn't work, and I'm using Meshroom, which works perfectly and is free. What shocks me the most is that people are still being duped. And that this programme is still here. And no, it's not slander, it's a fact, and I don't have to prove anything - I didn't make the software.

by the way, can I ask you to process a couple of these photos using meshroom.
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link

It would be great to compare whether meshroom can get a 3D object from two photos, or my program does it better.

I didn't say that Meshroom only did it with two photos. I just said that Meshroom was free and that at least it makes beautiful 3d captures. even if it takes many more photos, at least it makes real 3d renderings with texture .....

I completely agree that Meshroom is a great program.

The competitive advantage of my program is that it transfers the visual texture into the relief of the 3D model. This is cool. As a result, if we print a model even with plastic of one color, we get amazingly realistic 3D models:

https://ibb.co/Tcjzkxy

My program has a parameter "microrelief height". The default value is 0 mm, but if we set it to 2-3 mm, and process these 4 pairs of photos that I linked about, all the small details from the image will appear on the 3D model itself.

And another competitive advantage of my program: the ability to get 3D models from movies. Programs like Meshroom will not help here, since we cannot view an object from different sides. But in my program we can take two frames (if the camera in the movie moves relative to the object), and get a 3D model of what is in the movie (of course, only the relief of the visible part of the objects). I plan to prepare a large video with a dozen such examples of how I capture models from movies.

Dear Najmstar! Thank you for your interest in my program! And thanks for your constructive suggestion!

Most of my 'Amazing STL Creator' programs are very easy to use and ideal for beginner 3D printing people. But not this program.

This program requires 20 minutes to read the instructions and understand the principle of optical 3D scanning, and 5 minutes to watch a video about the program's capabilities. And your first 10 scans will be bad while you train. It can be difficult for newbies and people may write bad reviews. That's normal, I also don't like things that don't work right away)

However, this program should definitely be in the arsenal of every experienced 3D printer and everyone who professionally works with 3D models. As you can see, among the 623 who ed the program, only 12 people had negative reviews. I think more than 500 people are professionals and understand well that this program has fantastic capabilities:
1) this is the only technology available in the world that creates a 3D model from two photographs. In of software implementation, the program works perfectly, but it completely depends on what the s: to get high-quality models, you need to practice taking good pairs of photos.
2) take 2 photos of everything around with your smartphone at any time. And later (in a week?) you can run the program and process these pairs of photos and get 3D models
3) you can get 3D models from games and movies by simply taking two frames from there
4) completely local processing, without any Internet services. This program will remain with you forever, no one will disable the service. And you can scan intimate and confidential things.
5) you get a color scan in addition to the usual STL
6) you can scan distant objects, which no laser scanner allows: mountains, cityscapes, planets, galaxies
7) you can apply a microrelief to the main 3D object (turn information about color into microrelief) to further improve the detail of the object
8) the price of the program includes a full license for all STL files created by the program

To summarize:
1) I'm sorry, but I can't provide a trial version. Moreover, after some time I will increase the cost of the program to 20-30 USD, so that it will not be bought by beginner 3D printing people, but only by those who will seriously use it.
2) Please do not purchase this program unless you plan to leave a positive review and high rating.
3) I'm currently working on a version of this program for beginners. "Amazing STL Creator 12: 3D selphie". This program will allow people to one photo of a person's face and create a 3D model of the head from this photo using neural networks to restore the 3D relief.

So, please wait for the release of the new program, I plan for it to be released in a month.

Thanks anyway for your comment. I understand why you had such a logical question, and I would probably also ask the same thing)

About 10 years ago Autodesk released a wonderful mobile 3D scanner 123D. But after 2-3 years, Autodesk stopped ing this application.
https://www.youtube.com/watch?v=_6tmskUgGWE
https://www.youtube.com/watch?v=08kawwZgkOk
https://www.youtube.com/watch?v=3NwmQkApWv8
https://www.youtube.com/watch?v=B26Uo6oCGe8
https://www.youtube.com/watch?v=-jY2_EiIup4

The reason is very simple: a scanner is a measuring tool, and one need to use it carefully. Some s photographed objects not according to the instructions and complained that they received bad models. Autodesk decided that this application was causing reputational damage.

And now the 3D printing community does not have this wonderful 123D application(

I'm in the same situation. Some s do not try to understand the principle of shooting and do not look at the instructions. For example, some people confuse left and right photos and get a concave object instead of a convex one. Some people the same photo instead of right and left and fail to get 3D object. Some people's shooting points are too far apart and people end up with too high and unconnected relief (peaks)

This problem cannot be solved by improving the application (Autodesk had everything perfect, but we can’t get a good model from incorrectly taken photos). We need to somehow explain to people how to take two photographs correctly to get good scans. That's why I stopped at this stage: now I don't need to improve the application, but find a way to help people take 2 good photos.

I like the approach of Joseph Prusa, who specifically produced several videos to help people get good models using photogrammetry https://youtu.be/ye-C-OOFsX8?si=R8svoBIjWE2c_-R1&t=355

So I definitely recommend purchasing this app and spending a little time learning how. Moreover, this application has seven advantages over the Autodesk 123D application. Just look, recently, my colleague figured out the application and made his own bust: https://ibb.co/M9Q0M4Z

I plan two further developments in this direction:
https://www.youtube.com/watch?v=_6tmskUgGWE
1) make some very brief but very clear video, after which people will not make mistakes. And make a dozen videos where I scan about 100 various objects (and so that people can the original photos and also create these models)
2) soon I will release the "Amazing STL Creator 12: 3D selfie" application, which is aimed at creating a 3D model of the head from one photo. This is a specific application, for heads only, but it will not require people to master a more complex photogrammetric scanner.

Thank you for your anyway!

contrary to what the creator said, it never worked, he even refused to refund me on the pretext of an update that was never released. for me this is nothing more than a scam.

It also happened to me, I was never reimbursed, I feel cheated to this day.

Dear Diester!
So that you do not unfairly accuse me of fraud, I will be happy to refund you for this program.

Initially, I planned to improve the program, and I probably promised you this. But then I came across the fact that no improvements to the program will improve the processing if two photos ed by the do not match each other well. So I came to the conclusion that I need to make a lot of instructions - release a lot of new videos, where it will be explained in more detail and in a more interesting way how to use the scanner. So soon I will return to this direction.

As far as I understand, a month has ed during which I could return your money by clicking the "refund" button. What if I buy a couple of your models for 7 euros to compensate you for the costs of the program.

Dear Suzanuwo! Dear Diester!

I think it would be wrong of me to argue. Your mood is the most important thing. And I don't need to be right, but for people to be happy using my programs.

Usually I am always ready to return the money, focusing on the person's mood, but I understand that more than a month has ed and the "refund" on Cults will not work. But what if I offer you to choose several (3-4) of my other programs from the 17 programs that I have released? Please write which programs you want to receive.

Besides the 3D scanner, other programs are guaranteed to give results, since they do not require any training from the . It is enough to enter text or a JPG.

ASC1: Dual text illusion - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-create-dual-text-illusion
ASC2: Text spinner - https://cults3d.sitegames.net/en/3d-model/gadget/text-spinners
ASC3: Switchable business card - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-create-fantastic-business-cards
ASC4: Dual sided keychain - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-create-keychains
ASC5: Decoration - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-create-decorations
ASC6: Cookie cutter - https://cults3d.sitegames.net/en/3d-model/gadget/app-for-cookie-cutters
ASC7: Lithophany and bas-relief - https://cults3d.sitegames.net/en/3d-model/gadget/program-to-create-lithophany-and-bas-relief
ASC8: Bobble head - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-turn-a-photo-into-bobble-head
ASC9: 3D scanner - https://cults3d.sitegames.net/en/3d-model/gadget/3d-scanner-app-to-turn-2-photos-into-3d-model
ASC10: Flexy keychain - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-create-flexible-keychain
ASC11: Articulated - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-create-articulated-things
ASC15: String Art - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-create-string-art
ASC17: Through - https://cults3d.sitegames.net/en/3d-model/gadget/app-to-create--through-spirals

Dear Adanger! Thank you for ing the app!

Some people have problems because the software 3D scanner is a tool. And like any tool, it requires an understanding of the principle of operation and accuracy.

For example, some people confuse left and right photos and get a concave object instead of a convex one. Some people the same photo instead of right and left and fail to get 3D object. I like this video tutorial from Joseph Prusa where he discusses the features of the photogrammetric scanning method
https://youtu.be/ye-C-OOFsX8?si=R8svoBIjWE2c_-R1&t=355

Do I understand correctly that you had problems at the very beginning, with loading the JPG? This is strange, because usually all my 13 programs load the JPG normally. Could it be that this is an JPG with an extremely high resolution?

The photo of the man just gave a somewhat 3d stl so I'll try now with 150x150 and up the scan window to 20 and processing time to 10

also gave a STL, bit better,, be aware >>looks like program hangs but give it time ..it will proceed and stops when finished and saved the STL !!.. I will continue to experiment with it .It's not realy friendly but eventualy I wil get some results , not bad for the buck ;-) and I have something to do instead hanging out on the streets hahaha

Dear Scorioo42! Thank you for your experiments with the program!

The program tries to extract a maximum of two photos. But the photogrammetry method has many pitfalls. I like this video by Joseph Prusa where he describes the ins and outs of scanning through photos:
https://youtu.be/ye-C-OOFsX8?si=sFAxHVIEHv0FfF6j&t=363

If you experiment, it is better to experiment not with the program settings, but with how to take two photographs of an object differently. The program settings will change almost nothing if the two photos are not suitable.

This is a great program if practice using it. Need to acquire the skill of how to photograph an object so that two photographs carry information about its shape.

I use this program myself and would be happy to help you use it successfully too. If you send me your pairs of photos, I will be happy to try to process them and tell you what to change in the shooting conditions.

Dear Yasinacar! Thank you for ing my two programs!
Thank you for trying the 3D scanner! This is a wonderful tool, I use it myself and many of my friends use it to create models.

The advantage of this scanner is not that it easily produces perfect models. The point is that after some experience, the scanner allows to obtain unique scans that no other scanner in the world can make. I'll soon release a big video where I scan many more different objects, distant landscapes and pull out 3D objects from games, movies, YouTube videos, and stock stereo photos.

This software 3D scanner, like any tool, requires an understanding of the scanning principle and some practice. The specifics of this 3D scanner program are such that I cannot correct it using any software methods if the does not take two accurate photographs. To help people understand what two photos are needed - I have prepared four sets of photos and illustrated instructions in PDF.
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link
https://drive.google.com/file/d/15goxWi_ZoEB03AxvQ2zXat5VdKoVfwGv/view?usp=share_link

Here are the main mistakes in using the program, because of which people have difficulties. Perhaps your difficulties were caused by these same problems.

1) If people are photographing an object from two shooting points that are too far apart. Then it is difficult for the program to match the two images. When photographing a face, the distance between two shooting points should be no more than 5 cm.
2) Sometimes people confuse left and right photos and get a concave object instead of a convex one
3) If people view the STL file in the viewer, then on a one-color 3D model it is difficult to understand what exactly gives artifacts: an object, a background, or other objects. Therefore, it is better to view scans in color mode. If you drag both OBJ+MTL to a viewer ( https://3dviewer.net/ ), then you will see a colored object. Then it is immediately clear what to change in the shooting conditions.
4) Artifacts can appear on those parts of the object where the texture is poorly expressed, which makes it possible to find the same fragment of the object in two photographs. To avoid this, you can change the lighting conditions, or use colorful backgrounds and object coverage.

Please write how I can help you
Let us make this program work for you

Dear xXwinsXx! Thank you for trying this 3D scanner. This is a wonderful tool, I use it myself and many of my friends use it to create models.

This software 3D scanner, like any tool, requires an understanding of the scanning principle and some practice. The specifics of this 3D scanner program are such that I cannot correct it using any software methods if the does not take two accurate photographs. To help people understand what two photos are needed - I have prepared four sets of photos and illustrated instructions in PDF.
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link
https://drive.google.com/file/d/15goxWi_ZoEB03AxvQ2zXat5VdKoVfwGv/view?usp=share_link

It's great that you got good scans the first time! Only for some reason you mixed up the left and right photos, so you got a concave scan instead of a convex one. And you had too much distance between the shooting points (about 20 cm instead of 4 cm), so the scan turned out to be too elongated along the z-axis.

Dear Sheepgod! Thank you for ing the program and trying it out!

In fact, the dozens of scans that I received in my videos were obtained with the roughest settings of 100x100 or 150x150, and a processing time about 3 minutes. Longer settings only increase the quality quite a bit. The quality is much more influenced by the accuracy in obtaining the two original frames.

This stereo method of obtaining 3D models raises questions among many s of the program. So that a person can understand the principle and quickly master the program, I have prepared 4 pairs of photographs with which you can start
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link

I have also prepared illustrated instructions in PDF:
https://drive.google.com/file/d/15goxWi_ZoEB03AxvQ2zXat5VdKoVfwGv/view?usp=share_link

I will be happy to answer any questions you have about this program. I like this program, I use it myself, and I will be glad if you also receive a convenient and effective tool in the form of this program.

Dear Yann61! Thank you for downlading the program!

What do you mean by corrupted file? Do you receive a file in STL format?

Like any tool, this 3D scanner requires an understanding of the operating principle and a certain accuracy.

Please write what results you got from processing these four sample photographs?
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link

Here are the main mistakes in using the program, because of which people have difficulties. Perhaps your difficulties were caused by these same problems.

1) If people are photographing an object from two shooting points that are too far apart. Then it is difficult for the program to match the two images. When photographing a face, the distance between two shooting points should be no more than 5 cm.
2) Sometimes people confuse left and right photos and get a concave object instead of a convex one
3) If people view the STL file in the viewer, then on a one-color 3D model it is difficult to understand what exactly gives artifacts: an object, a background, or other objects. Therefore, it is better to view scans in color mode (OBJ+MTL files), then it is immediately clear what to change in the shooting conditions.
4) Artifacts can appear on those parts of the object where the texture is poorly expressed, which makes it possible to find the same fragment of the object in two photographs. To avoid this, you can change the lighting conditions, or use colorful backgrounds and object coverage.

Please write how I can help you
Let us make this program work for you

Dear 1143! Thank you for experimeniting with the program!

You write that these photo samples, which I recommended to start with, are not processed.
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link

But I have just processed these 4 samples with the program, with the default program parameters. Look, please, everything turns out fine (by default, there is a low resolution of 100x100):
https://ibb.co/PhPFWB9
https://ibb.co/g76Nd4B
https://ibb.co/Pmm0f6g
https://ibb.co/QQngLJp

If you get not exactly the same models, please send your screenshots (you can also them to imgbb.com or send them to my email [email protected]). If your program does not produce exactly the same 3d models, this would be very strange, and this needs to be dealt with.

My other programs (and there are already 12 programs in total) are simpler: they simply convert a photo or text into STL. Therefore, their results are more predictable. So far, about 5,000 people have ed my programs about 13,000 times. And in most cases, people are quite satisfied with the programs.

So let's make this program work for you

same here :)

Dear Zefirosaryanov! Thank you very much for ing two my programs!

I really love this program and I will be glad to help you in any way to get good scans with it. You can send photos ([email protected]) and I will be happy to process them or tell you what to change in the shooting conditions

A software 3D scanner, like any tool, requires a little understanding of how it works and some practice.

In order for people to quickly master this scanning technology, I specifically provided sample photos and illustrated instructions.

Of the approximately 1,000 people who have ed this program (from 3 platforms), approximately 900 people are successfully using it. This software 3D scanner has many advantages over other similar programs, from completely local processing (allowing confidential and intimate scans) to the ability to get 3D models from movies, games, and the ability to apply micro-relief for 3D printing.

Here are the main mistakes in using the program, because of which people have difficulties. Perhaps your difficulties were caused by these same problems.

1) If people are photographing an object from two shooting points that are too far apart. Then it is difficult for the program to match the two images. When photographing a face, the distance between two shooting points should be no more than 5 cm.
2) Sometimes people confuse left and right photos and get a concave object instead of a convex one
3) If people view the STL file in the viewer, then on a one-color 3D model it is difficult to understand what exactly gives artifacts: an object, a background, or other objects. Therefore, it is better to view scans in color mode (OBJ+MTL files), then it is immediately clear what to change in the shooting conditions.
4) Artifacts can appear on those parts of the object where the texture is poorly expressed, which makes it possible to find the same fragment of the object in two photographs. To avoid this, you can change the lighting conditions, or use colorful backgrounds and object coverage.

Thank you for your !
Please write how I can help you
Let us make this program work for you

nice .. ok .. send me a link for with 3 pictures : 2 for model and 1 with information " click here .. click here" ... and put it in the description of the so there will be no situation like this

Dear Tkjeldsen! Thank you for ing the program! Thank you for appreciating the idea!

My programs are developed in the Lazarus environment, and have a simple structure and functions. Therefore, they usually work well under different versions of Windows and different processors. Only once a person complained to me: he had some kind of rare processor (not AMD) and a 32-bit operating system.

Write me, please, what kind of message appears, after which the programs do not work?

Can you get such a standard Windows warning that you are running an EXE file? (this is Windows 10).
https://ibb.co/mGBWJ8g
In this case, we need to choose "More info" and then "Run anyway".

Let's try to solve this problem. Thank you for your !

ok program is running..... somehow it did't have premisions to run... thx for a great program.. is it by the way multi treaded , have 32 cores to play with

Thank you for your high rating! 32 cores is very cool. I sometimes write multi-threaded programs. But I specifically made the programs of the "Amazing STL Creator" series with a very simple structure, single-threaded, so that it would work reliably for people with a variety of computers. And so, it could work 32 times faster!

Dear Leeullah! Thank you very much for using 5 of my programs!

If you send me a couple of your photos by email ([email protected]), I will be happy to analyze them and tell you how to get good 3d models.

With this method of 3D scanning (stereo vision), 3-4 rules must be taken into when taking photos.

1) Please start by processing these four examples to make sure the program works and creates good 3d models:
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link

2) Please view your models created by the program in the color viewer (not in the black and white viewer) https://3dviewer.net/ . You need to drag into this viewer both OBJ and MTL files that the program creates. A color image will allow you to see what is good in the 3D model and what is bad (object or background). This will allow you to immediately understand how to change the conditions of photography in order to get a good model.

3) Common mistake: two shooting points are too far apart. The distance between the two shooting points should be small (5-8 cm) when shooting an object at a distance of 50-100 cm. The two photos should look almost the same.

4) Please check that there are no shiny, translucent or solid color areas on the object and background. Such areas do not contain information about the volume and no program will be able to restore the 3D model. Therefore, it is desirable to use a chaotic motley background. And the shiny areas on the scanned object can be somehow textured (sprinkled with flour, chaotically colored, pasted over with patterned ribbons, put many pieces of paper 1x1 mm on them)

In addition, in about a week I will make an improvement to this program, aimed at heuristically removing random peaks and smoothing the surface. Then, even if s make mistakes, they will still get a pretty similar 3d object. You will receive a message from Cults that you can the new version of the program for free.

Thank you for your !

Dear Shaf2! Thank you for ing the program!

With this method of 3D scanning (stereo vision), 3-4 rules must be taken into when taking photos. The program works well, but photos should help it work.

1) Please start by processing these four examples to make sure the program works and creates good 3d models:
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link

2) Please view your models created by the program in the color viewer (not in the black and white viewer) https://3dviewer.net/ . You need to drag into this viewer both OBJ and MTL files that the program creates. A color image will allow you to see what is good in the 3D model and what is bad (object or background). This will allow you to immediately understand how to change the conditions of photography in order to get a good model.

3) Common mistake: two shooting points are too far apart. The distance between the two shooting points should be small (5-8 cm) when shooting an object at a distance of 50-100 cm. The two photos should look almost the same.

4) Please check that there are no shiny, translucent or solid color areas on the object and background. Such areas do not contain information about the volume and no program will be able to restore the 3D model. Therefore, it is desirable to use a chaotic motley background. And the shiny areas on the scanned object can be somehow textured (sprinkled with flour, chaotically colored, pasted over with patterned ribbons, put many pieces of paper 1x1 mm on them)

5) Sometimes people mix up the left and right photo and get a concave object, not a convex one.

These are the basic rules that allow you to get good 3d models.

Please email me if you continue to have problems. If you send me a couple of your photos by email ([email protected]), I will be happy to analyze them and tell you how to get good 3d models.

Thank you for your !

Das gleiche ist bei mir auch, schrott Software!!

Dear Arminlimberger! Thank you for ing the program!

1) Have you tried processing these sample photos?
https://drive.google.com/file/d/14bQn_mxSkTIw3Jl3Tgh-aZcu78TXxC4l/view?usp=share_link
These photos give good 3d models. If you take photos of your object by the same way, you also get good 3D models.

2) Have you tried to view the 3D models in color viewer? Color allows to immediately understand which part of the object or background gives artifacts (peaks) on the 3D model, and correct the conditions for taking photos.

Here is the color viewer https://3dviewer.net/ . You need to drag into this viewer both OBJ and MTL files that the program creates.

3) Most common scanning mistake: two shooting points are too far apart. The distance between the two shooting points should be small (5-8 cm) when shooting an object at a distance of 50-100 cm. The two photos should look almost the same.

Please email me ([email protected]) if you continue to have problems. If you send me a couple of your photos by email, I will be happy to analyze them and tell you how to get good 3d models.

Dear Kllaus! Thank you for ing the program!

The application is not blocked. There may be a problem that the debug graphics stop updating, but the application continues to work in the same way, and as it is written on the button, and after 1-2 minutes it writes files with results. The graphics stop updating if you additionally press something on the program interface, or some other program intercepts the focus. I planned to understand and modify the program so that the debug graphics would not disappear, but now this does not interfere with the full operation of the program.

From the experience of recent days, when people get bad scans - most often the problem is as follows:
1) confuse the left and right photos in places and get a concave object instead of a convex one
2) distance between two shooting points is too long (should be 5-8 cm when shooting an object at a distance of 30-100 cm)
3) the object or background is not textured enough and the program cannot find the same fragments (to exclude this, you need to use something with a chaotic coloring instead of the background and choose or make the object a non-monotonous color and non-reflective)

In any case, on sample photos, the program should give good enough scans, as far as such a measurement principle allows. If the sample photos turn out to be bad scans - please send me to [email protected], it would mean that the program is doing something wrong.

In addition, it is desirable to view the results first of all in a color viewer https://3dviewer.net/
You need to drag there both OBJ and MTL files at the same time. And on a colored 3D object, you can immediately see: the problem is with the background or object, and what the problem.

Can I ask you to consider these suggestions and try to process at least a sample set of photos.

Thank you very much for your !