|
|
|||||
|
|
|||||
|
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
|
