|
|
|||||
|
|
|||||
|
SHAllow Looks One Map (SHALOM™) is a method of REAL
TIME OBJECTS DETECTION that uses SHALLOW NEURAL CLASSIFIERS and CELLULAR
AUTOMATA in order to identify objects in the frames of a video looking only
one time at any single frame. SHALOM™ is inspired by YOLO algorithm but it is based on
Shallow Neural Networks and Cellular Automata. SHALOM™ has been designed to
speed up the identification of specific objects and determine their exact
position. SHALOM™ works with high speed of execution both in the
"features extraction" phase that uses a single image scan without
ROS (Region Of Scanning) and ROI (Region Of Interest), and in the pattern
recognition phase with Shallow Neural Networks on SIMD processors or Neuromem®. Mythos™ technology
enables SHALOM™ to run on Von Neumann processors like BAE SYSTEMS RAD750™.
The Cellular Automaton manages the behaviour of the fixed grid. |
||
|
|
|
|
|
|
|
©2024_Luca_Marchese_All_Rights_Reserved Aerospace_&_Defence_Machine_Learning_Company VAT:_IT0267070992 NATO_CAGE_CODE:_AK845 Email:_luca.marchese@synaptics.org |
|
Contacts_and_Social_Media
|
