Object Detection and Recognition using Knowledge Growing System

Abstract

One of challenging applications for Artificial Intelligence (AI) is image detection and recognition. Several years back, various techniques in deep learning have been explored to obtain the maximal result. Convolution Neural Networks (CNN), a kind of deep learning architecture and others have been used for such challenge. To build a deep learning model closest to 100% classification accuracy requires high computational task, hugh effort to label the dataset, and the network has to be trained with a huge number of data before used. Motivating by these considerations, we would like to enhance the capability of Knowledge Growing System (KGS) for image detection and recognition task. KGS is defined as a system that has capability to grow or develop its own knowledge along with the accretion of information as time passes. It is emulating human brains capability in developing knowledge by performing information fusion. KGS emulates how data and information are fused within the brain. KGS becomes intelligent by developing its own knowledge from new data and information perceived by its sensory systems. It does not need to be trained before used, does not need huge data just that one it interacts with, and is very light. By enhancing its capability in growing the knowledge by performing learning by interaction and use the grown knowledge, KGS is able to detect and recognize an object through its image much faster than existing AI methods.

Document Details

Document Type
DoD Grant Award
Publication Date
May 05, 2021
Source ID
N629092112017

Entities

People

  • Arwin Sumari

Organizations

  • Office of Naval Research
  • State Polytechnic of Malang
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Exercise and Sports Science.
  • Neural Network Machine Learning.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • AI & ML - Neural Networks