Today I've released Imaginario 0.9. The big feature coming with this new release is a face tagging flow which I believe will be the fastest and simplest you've ever used, despite it being all manual. I even sat down and spent some quality time with Blender to prepare a video to show it off:

While some people might actually think that I spent more time for making the video than for implementing the face tagging feature itself, this couldn't be farther from the truth: the face tagging branch has been being worked on for at least three months (of course, that's my spare time — so it's actually less than one hour per day) and consisted of more than 40 commits (after squashing all the fixups), whereas for the video I spent no more than a couple of hours.

I would appreciate if the curious could go and try it out, and let me know about any issues you should find: there are built packages for Linux (AppImage), macOS and Windows. I do also have an Ubuntu PPA where nightly images are built, but I'm not sure if I can recommend that one, since I've not been using it myself and have no idea whether those packages actually even start. But you are welcome to try :-)

Your feedback will help me do better, so please don't be shy!

Lets say you have mesh data in the typical format, triangulated, vertex buffer and index buffer. E. g. something like

>>> vertices

[[[ 46.27500153  19.2329998   48.5       ]]

 [[  7.12050009  15.28199959  59.59049988]]

 [[ 32.70849991  29.56100082  45.72949982]]

 ..., 

>>> indices

[[1068 1646 1577]
 [1057  908  938]
 [ 420 1175  237]
 ..., 

Typically you would need to feed it into OpenGL to get an image out of it. However, there are occasions when setting up OpenGL would be too much hassle or when you deliberately want to render on the CPU.

In this case we can use the OpenCV to do the rendering in two function calls as:

img = np.full((720, 1280, 3), 64, dtype=np.uint8)

pts2d = cv2.projectPoints(vertices, rot, trans, K, None)[0].astype(int)
cv2.polylines(img, pts2d[indices], True, (255, 255, 255), 1, cv2.LINE_AA)

• Wire-frame rendering

See the documentation of cv2.projectPoints for the meaning of the parameters.

Note how we only project each vertex once and only apply the mesh topology afterwards. Here, we just use the numpy advanced indexing as pts2d[indices] to perform the expansion.

This is pretty fast as well. The code above only takes about 9ms on my machine.

In case you want filled polygons, this is pretty easy as well

for face in indices:
    cv2.fillConvexPoly(img, pts2d[face], (64, 64, 192))

• Filled polygon rendering

However, as we need to a python loop in this case and also have quite some overdraw, it is considerable slower at 20ms.

Of course you can also combine both to get an image like in the post title.

From here on you can continue to go crazy and compute face normals to do culling and shading.

• But you can only get that far without a depth buffer

So after having made fun of people for “wearing toothbrushes”, I finally came to buy such headphones for myself.

I just released PhotoTeleport 0.12, which includes the feature mentioned in the title of this blog post. Given that it took me some time to understand how this could work with Qt, I think it might be worth spending a couple of lines about how to implement it.

Recently I've grown an interest to the indieweb: as big corporations are trying to dictate the way we live our digital life, I'm feeling the need to take a break from at least some of them and getting somehow more control over the technologies I use.

Some projects have been born which are very helpful with that (one above all: NextCloud), but there are also many older technologies which enable us to live the internet as a free distributed network with no owners: I'm referring here to protocols such as HTTP, IMAP, RSS, which I perceive to be under threat of being pushed aside in favor of newer, more convenient, but also more oppressive solutions.

Anyway. The indieweb community is promoting the empowerment of users, by teaching them how to regain control of their online presence: this pivots arund having one's own domain and use self-hosted or federated solutions as much as possible.

One of the lesser known technologies (yet widely used in the indieweb community) is webmentions: in simple terms, it's a way to reply to other people's blog posts by writing a reply in your own blog, and have it shown also on the original article you are replying to. The protocol behind this feature is an recommendation approved by the W3C, and it's actually one of the simplest protocol to implement. So, why not give it a try?

I already added support for comments in my blog (statically generated with Nikola) by deploying Isso, a self-hosted commenting system which can even run as a FastCGI application (hence, it can be deployed in a shared hosting with no support for long-running processes) — so I was looking for a solution to somehow convert webmentions into comments, in order hot to have to deal with two different commenting systems.

As expected, there was no ready solution for this; so I sat down and hacked up Bussator, a WSGI application which implements a webmention receiver and publishes the reply posts as Isso comments. The project is extensible, and Isso is only one of the possible commenting systems; sure, at the moment it's indeed the only one available, but there's no reason why a plugin for Static Man, Commento, Remark or others couldn't be written. I'll happily accept merge requests, don't be shy — or I can write it myself, if you convince me to (a nice Lego box would make me do anything

We live in strange times. People are so filled with hatred and prejudices that their brain becomes unable to parse the simplest sentences. I take this issue to heart, because it could happen to anyone — it has happened to me before (luckily, only in private online conversations), where an acquaintance of mine accused me of saying things I never said. And it happens to famous people all the time. Guys, just because you hate person X, you should not skip over parts of their speech or suppress context in order to make it look like they said something terrible or stupid, when they didn't.

Now it happend to Richard Stallman, with a whole wave of hateful people accusing him of saying something that he didn't say. Let's start with the VICE article, titled "Famed Computer Scientist Richard Stallman Described Epstein Victims As 'Entirely Willing'", which insists in quoting only two words out of Stallman's sentence:

Early in the thread, Stallman insists that the “most plausible scenario” is that Epstein’s underage victims were “entirely willing” while being trafficked.

Except that he didn't say that. Why not quote the whole sentence? It's not such a long sentence, really! Just follow the link to the source, which provides a complete excerpt of Stallman's words:

We can imagine many scenarios, but the most plausible scenario is that she presented herself to him as entirely willing. Assuming she was being coerced by Epstein, he would have had every reason to tell her to conceal that from most of his associates.

Now, English is not my native language, but I read it well enough to understand that “to present oneself as” and “to be” are different expressions having very different meanings (and, in most context, actually opposite ones!). You don't need to be Shakespeare to understand that. You only need to either stop hating or, if you really cannot help it, at least stop projecting your prejudices onto the people you hate. Hate makes you blind.

It's sad to see otherwise intelligent people take stupid decisions because of such misunderstandings.

I for one, stand in solidarity with Richard Stallman and with the English language.

(please note that this is not an endorsement of everything Stallman might have said in the past; I don't follow him that closely, and it may be that he also happened to say terrible things in this very thread. I'm only commenting this very specific issue, and I know that in this very specific issue he's being wrongly accused)

The wonders of improvised Mastodon instances: one node disappears after an outage caused by a summer heatwave, leaving its users no way to migrate their data or to notify their followers.

After about one month of waiting for the node to come up or give some signals of life, I've decided to create a new account on another instance. If you use Mastodon and you were following me, please forgive me for the annoyance and follow me again here.

Camera calibration just got even easier now. The pose calibration algorithm mentioned here is available as web-service now.

This means that calibration is no longer restricted to a Linux PC – you can also calibrate cameras attached to Windows/ OSX and even mobile phones.
Furthermore you will not have to calibrate at all if your device is already known to the service.
The underlying algorithm ensures that the obtained calibrations are reliable and thus can be shared between devices of the same series.

Aggregating calibrations while providing on-the-fly calibrations for unknown devices form the calibDB web-service.

In the future we will make our REST API public so you can transparently retrieve calibrations for use with your computer vision algorithms.
This will make them accessible to a variety of devices, without you having to worry about the calibration data.

When using Nextcloud it makes some sense to host it yourself at home to get the maximum benefit of having your own cloud.

I was looking at TripAdvisor, some days ago. It's a very useful site, filled with user-generated advice and reviews, which has become almost a must for travellers. But I never like it when a private entity gets so much power over our lives (even if it's — currently — exercised in total fairness).

I would like to have a federated TripAdvisor-like network. I duckduckwent for a while, but I didn't find anything similar (if you know of some project of that kind, please let me know in the comments).

But thinking more about the issue, I realised that I probably wouldn't even bother to type my reviews into that site; I have a blog, so ideally I would like to have the option to write my review here, and then have the site import it. Technically, it could work with a webhook, or even a periodic check (real-time updates would not be a requirement here) over a URL I've linked to in my profile on that site. Then, in order for the posts to be imported, they would have to be entered in a standard format: maybe some keywords (or invisible HTML elements) could be used as markers for the content that need to be extracted from an otherwise ordinary blog post, or the relevant content could be replicated in a different format in the HTML headers (this, though, would require some additional work).

And while we are at it, why not extend this to other social networks? I use Mastodon, for example, and occasionally I send out a toot with a link to my latest blog post in there. But it would be much nicer if I could somehow set a special mark into my posts while composing them, to have them automatically tooted out on my account (this could probably be implemented as a standalone service, authorized to act on my Mastodon account — similarly to how the Mastodon-Twitter crossposter app works).

But I'm rather confident that I'm not the only one having this kind of needs, and that's why I'm writing this blog post: maybe someone out there has already found a solution, or has some more concrete ideas? If so, I'm all eyes!