We've released a bug-fix only firmware update (1.2.3) which addresses the following issues:
• Second drive might not be recognized if the first drive was added with v1.1.17 or older.
• Unit might get stuck analyzing content if the unit has changed its drive.
• Locations could not be analyzed properly in some cases.
An iOS update (1.5.2) is also available which fixes the crashing issues some users have.
Thank you all,
The Monument Team
@Patrick; You are right of course, and our apologies as juggling bug fixes and implementing features with limited start-up resources always seem to beat our best intentions. We did get everything more or less ready and in final testing, albeit far later than we wanted to, submitting to iOS store Friday afternoon. We are indeed easily 2 weeks behind that promise. Our sincere apologies,
Bug fix didn't work.
Unit was stuck processing 7025 photos.
It moved on then got stuck processing 5025.
The blinking LED is driving me crazy, just like the monument.
It is just going by what's in the EXIF of the photo for camera name. As you zoom into countries, the cities may spilt into other ones as well.
General Problem Resolution After version 2.0.0 (app) / 1.3.0 (Firmware) Update:
1- Update your App first before updating the firmware. If you don't update your App, communication will be impossible, as we changed protocol.
2- If you updated the Monument, you can verify firmware version by looking at the HDMI output on a TV/Monitor, you should see it at the lower left corner (1.3.0).
3- If you successfully updated the firmware but now cannot connect after updating the App, delete the App, and reinstall. You do not need to reset your Monument.
If you have done these and still cannot connect, or if you have persistent RED or YELLOW LED on your Monument, please file a ticket.
how to upgrade it ?
i have now 152 & 122
@rigaux: updates will appear in your mobile device.
Ok, we all get that facial recognition is not what you want it to be, or what it's supposed to be.
That aside, let's see if we can help you, but helping US understand how to teach the software.
Let's use this first example to start:
1. I don't have a face or name setup for my friend Scott. But I see him in the list.
I click on his face, name him and save. But when I go to confirm that all the other faces are him, I notice that in fact, more are of ME, that of him. Only the first image is him. The rest are others but MOSTLY, me. Should I a) Remove ALL other images, and leave the one that was on the face page of Scott or b) remove everything except the most frequent face (in this case me) then save, and rename it to me? The reason I ask, is because when I did the latter, it still showed Scott's face when the merge occurred. I don't want to be doing anything that's going to counter act the learning algorithm.
2. There are a bunch of faces, that are from people I don't really speak to anymore, or it's hard to tell who it is, or it's not a real person (newspaper, comic, cartoon, etc) or you simply don't want that person saved. (Plus, you still can't preview BEFORE you save. We REALLY need that feature if we're going to be able to help you teach the algorithm) should we delete them? What exactly does it mean, when we delete a face from the new face screen? Are we remove what you think that GROUP of people are? Are we removing just that specific face you displayed there that you thought was a face from being any future faces?
3. I've been TRYING to teach it faces, but I'm noticing that each time I select a face to add, it has about 2 or 3 faces to add or merge of that actual person, and then the SAME photos keep popping up. I remove them, get it down to just the person I'm working on, and then move on to add the next new face, only to realize that the same photos I've JUST removed, are now in THIS face group. How is this possible. If I have 10 faces for person 1, and only 1 of them is actually them, then I go to the next person, who also has 10 faces, and 8 of them are from the previous face. But it's not like the number increased. If it was making a correction on the fly, and you saw the numbers go up, ok, that's one thing. But they aren't. What's going on here.
4. Can we get a better list? Maybe some filter options? Rather than just by number of photos per face? And not a bar. So we can see more than like 5 faces at a time.
5. There has to be SOME kind of strategy you have in mind for US to use, to help you teach the algorithm. I may not have asked all the right questions, but you get what kind of things I'm trying to learn right? I have pictures that are really really dark, but you can KIND of make out a face. Should we remove that? What about photos where I know it's a face, but really, it's a silhouette and not really a face, but it IS that person. Should we save it? Is telling the algorithm that "Yes, this IS that person (even if it's not there face)" incorrect or is it ok to teach the algorithm that based on other key features in the photos, we deduce that person x is in this photo?
At least this way, if I KNOW how you want us to teach your software, I'm not being counter productive by getting it to just sort things the way I want them.
Thanks and look forward to hearing what you can tell me.