I have been suffering from eyestrain since 2019, at first only when I was in front of a screen for a long time, but today things are worse.
With this experience and some curiosity, I want to understand how my body work. Eyestrain is well documented and I read a lot about.
Following informations and theories will be documented as good as I can, since I am not a Doctor, a Specialist or a Scientific. I hope that my work can give me some complementary data about my eyes health.
I can be wrong de but en blanc, but anyways it will be a way to learn OpenCV and to understand statistics.
I want to get enough data about my eyes during my work on the computer to understand what can make my eyestrain appear. So that I can prevent it.
Theses metrics will be compared to average metrics and analysed so that I can find if I doesn't blink enough, or, if I blink enough maybe that will put ligth on another problem, like bad quality tears.
This study contains lot of interesting informations.
Starting by the introduction we can read following lines:
In addition to a reduction in the number of blinks, an increase in the number of incomplete blinks during computer use has also been reported.18
Results of the study :
The mean ± standard deviation blink rate was 19.74 ± 9.12/min at baseline. The blink rate decreased significantly under both reading conditions (to 11.35 ± 10.20 and 14.93 ± 10.90/min when reading from a book and a tablet, respectively).
We can notice in the conclusion that a notion of "frequent blinkers" exists.
The mean spontaneous baseline blink rate was 19.74 ± 9.12/min, which is slightly higher than the values reported in recent studies. [...] The reason could be that, in our current study, we did not exclude subjects who might be classified as frequent blinkers (blink rate of >21/min).
This study doesn't take into consideration this notion. So the first question (Q.1) is: what is representative of "mean blink rate"? Should "frequent blinkers" be taken in consideration?
Study 2 (S.2): "Analysis of blink rate patterns in normal subjects"
Mean BR at rest was 17 blinks/min, during conversation it increased to 26, and it was as low as 4.5 while reading.
Extending the Q.1 : Can we etablish that in normal conditions human blink their eyes around 18 times per minute?
The calculation of the blink rate becomes more accurate over time (Left graph).
In my case, with the actual blink tracker, I get a final measure of 22.6 blinks/min. It gives an answer to the Q.1 : If I am a "frequent blinker" (blink rate of >21/min), the first study may be a good base. I say "if" because the tracker is not quite accurate for the moment.
During tests I noticed that following parameters may affect image recognition :
- P.1 : Not enough light
- P.2 : Bad frequency of lights (cause a )
- P.3 : Glasses
- P.4 : Other people's face in background
I used this post of Stepan Filonov as base to start the project.