Problem with HUSKYLENS Object Classification function
GeorgeLambev 2022-08-01 18:04:37 1168 Views7 Replies Hi all! I am trying to use the Object Classification function and have a slight issue.
Problem:
Learnt Objects ID1 = bottle, ID2 = keys. (or whatever else could be the object, tried it multiple times on different objects)
No issues when identifying and distinguishing them. However, no matter where the camera is pointing to, it always “identifies” ID1 or ID2 even though they are not present at the current frame.There is constantly square with either ID1 or ID2 on the screen, switching between the two of them. It's like “I don't know what is this, but if I have to guess I'd say it looks closer to ID1”.
What I want to achieve:
Show ID1 or ID2 ONLY WHEN such object is recognized. If there is no such object in the frame, then not showing any square with ID at all.
I was looking for a setting to set the “confidence level”. For example “Show ID when there is >80% certainty this is the object”.
Any feedback will be greatly appreciated!
Hello George,
You might consider identifying the background as the default return.
Tonny12138 Hi and thank you for your reply! This is a good suggestion but unfortunately not an option in this case. Background is dynamic.
In fact, you can press the study button long enough to learn from multiple angles while the subject is learning, and ultimately improve the effectiveness of the training. In addition, you can also change the recognition threshold in the Settings.
Winster Hi and thank you for your answer! I have indeed long pressed the study button and given the camera different perspectives of the objects. It can recognize them flawlessly. Problem is that when none of them are present in the current frame, camera is still constantly “identifying” ID1 or ID2. As far as recognition threshold, can you please navigate me? I am using the latest firmware 0.5.1.aNorm and not finding it. Neither in General Settings nor Object Classification Settings.
up
Hi and thank you for your reply. Both good to know I am not the only one and sorry you are having problems as well. It is very pity that what could have been a really great product falls short in some key moments. I will have to scratch Huskylens from my project and hope for better days/firmware updates. Apparently, for now it is just a toy to play with.
GeorgeLambev Hello George,
Just replying in support of the problem you have described.
I am experiencing the same thing and other anomalies with the way objects are handled.
I tried documenting them here in hopes of an answer; however, the support staff is totally unresponsive
to HuskyLens issues. It is like they dumped the product and don't care about it anymore.
The current firmware level and its performance details leaves much to be desired for any meaningful application.
tguneysu 
