|
Machine/Deep Learning |
|
Machine or Deep learning are different techniques that is parts of the AI description.
These techniques are used to interpret the data from different sensors like traditional cameras as well as other types such as thermal cameras, Laser, Lidar, ultra sonic and other types of sensors.
Together with our partners we offer this in various areas, from finding objects with cameras to interpret accelerometers to differentiate a normal movement versus falling.
What this can be used to achieve is still unexplored, for sure is that the ones starting to use the technique will create themselves a advantage towards their competitors and find new values to offer their customers.
Below a simple description of pros and cons of the different methods:
These techniques are used to interpret the data from different sensors like traditional cameras as well as other types such as thermal cameras, Laser, Lidar, ultra sonic and other types of sensors.
Together with our partners we offer this in various areas, from finding objects with cameras to interpret accelerometers to differentiate a normal movement versus falling.
What this can be used to achieve is still unexplored, for sure is that the ones starting to use the technique will create themselves a advantage towards their competitors and find new values to offer their customers.
Below a simple description of pros and cons of the different methods:
|
Machine learning
+ Simpler and less computational power + Good results with small data sets + Quick to train a model - Accuracy plateaus - Need to try different features - Need to try different classifiers to achieve best result |
Deep learning
+ Learns classifiers automatically + Accuracy is extremely good + Learns features automatically - Requires large data sets - Computationally intensive |
