Proposed Improvement for Auto-Brightness in Mobile Devices.
DARSHIT PANDYA
Smartphones, as we know, have become a inseperable part of human life. As the days pass, the influence of this mobile device on humans
will increase exponentially to what it is today. Although Smartphones are typically Smart enough, they do lack smartness in some
aspects.
One aspect where the smartphone lacks is - Auto-Brightness Adjustment. Although it seems to be present as a feature across almost
all mobile devices, it is still not reliable enough. I believe almost all of the smartphone users would have experienced issues with this
feature. To be clear, this feature, sometimes, does opposite to what it should do. For example, even if there is considerable light present,
it will still decrease the Brightness; opposite to increasing it for better sight, and vice-versa.
This problem could be understood by considering a practical scenario. Assuming a person is holding a smartphone in between the source of light
and the eyes (See Figure below). Now, as the light sensor is on the front side, it detects less light than what the actual scenarion is. FYI,
the light sensor is used to measure and estimate the light w.r.t. the eye of the user. Accordingly, as it detects less intensity of light than
the acual light, it starts to decrease the brightness.
For resolving this issue, I am proposing a new method. Briefly, this method uses the images from both - front & rear cameras, and supplies it
to the trained Machine-Learning model; which predicts the approximate intensity of light and the value by which the brightness should be increased/decreased.
This Machine Learning model is trained with such images and their relative light intensities. FYI, the actual vale of the light sensor will also be
utilized as there are some outlier situations where light sensors become essential.
(This work is currently under construction and will be soon available publicly.)
Hope you liked the blog! If so, connect with me on FB/Insta/Twitter/LinkedIn by following the links in the about section.
~ Darshit Pandya
