Wake up, Neo! No spoons.
Recently, I paid quite a lot of attention to studying Flash applications with the aspect of using 3D objects, and in particular the so-called augmented reality . Curiosity and surprise from what he saw, as often happens, quickly turned into a desire to create something similar. It was boring to create cloned applications with already implemented ideas, and generally “not comme il faut”, therefore it was decided to invent something new.
The main disadvantage of existing Flash applications of "augmented reality" is the mandatory presence of a specific marker. Of course, you can print or draw it, but for practical use it is very inconvenient. And what if to make a “marker” not a square on a piece of paper, but yourself? In the world for quite a long time there is a well-established technology for recognizing human faces in an image, and I decided that you could use an observer's face as a marker.
After some searches, I found the Marilena library on the ActionScript 3.0 network , based on the above-mentioned OpenCV technology (by the way, it appeared as AS3 relatively recently). Of course, it has its drawbacks, in particular, it consumes quite a lot of resources for searching, and also in its original form does not take into account the rotation angles of the search object. Nevertheless, she copes with the task of finding a person in an image, and does it perfectly well.
')
So, having all the tools in hand, all that remained was to come up with an idea and it came by itself. A few days later this project was born:
Wake up, Neo!
I must say that in order to feel in the role of Neo, you will need a webcam and a relatively powerful computer. Unfortunately, it was not possible to test on different machines, perhaps you report significant problems in the comments :) However, I think this is the case when it is better to try it yourself.
In fact, this is not exactly “augmented reality” in the form in which it appeared to us earlier, but another way of interaction between the human interface and the computer interface. Yes, and practical application, which I spoke at the beginning, this toy is not much - yet the resource intensity of the search makes itself felt. However, it seems to me that the use of such technologies in the future may (and certainly should) affect the construction of user interfaces.
Enjoy!
PS The archive with the source code and a brief description are presented on this page .
UPD: Today I discovered an optimized, much faster version of the face detection library, which opens up new perspectives.
The main disadvantage of existing Flash applications of "augmented reality" is the mandatory presence of a specific marker. Of course, you can print or draw it, but for practical use it is very inconvenient. And what if to make a “marker” not a square on a piece of paper, but yourself? In the world for quite a long time there is a well-established technology for recognizing human faces in an image, and I decided that you could use an observer's face as a marker.
After some searches, I found the Marilena library on the ActionScript 3.0 network , based on the above-mentioned OpenCV technology (by the way, it appeared as AS3 relatively recently). Of course, it has its drawbacks, in particular, it consumes quite a lot of resources for searching, and also in its original form does not take into account the rotation angles of the search object. Nevertheless, she copes with the task of finding a person in an image, and does it perfectly well.
')
So, having all the tools in hand, all that remained was to come up with an idea and it came by itself. A few days later this project was born:
Wake up, Neo!
I must say that in order to feel in the role of Neo, you will need a webcam and a relatively powerful computer. Unfortunately, it was not possible to test on different machines, perhaps you report significant problems in the comments :) However, I think this is the case when it is better to try it yourself.
In fact, this is not exactly “augmented reality” in the form in which it appeared to us earlier, but another way of interaction between the human interface and the computer interface. Yes, and practical application, which I spoke at the beginning, this toy is not much - yet the resource intensity of the search makes itself felt. However, it seems to me that the use of such technologies in the future may (and certainly should) affect the construction of user interfaces.
Enjoy!
PS The archive with the source code and a brief description are presented on this page .
UPD: Today I discovered an optimized, much faster version of the face detection library, which opens up new perspectives.
Source: https://habr.com/ru/post/55001/
All Articles