SPRADC9
july 2023
AM62A3
,
AM62A7
1
Abstract
Trademarks
1
Introduction
1.1
Defect Detection Demo Summary
1.2
AM62A Processor
1.3
Defect Detection Systems
1.4
Conventional Machine Vision vs Deep Learning
2
Data Set Preparation
2.1
Test Samples
2.2
Data Collection
2.3
Data Annotation
2.4
Data Augmentation
3
Model Selection and Training
3.1
Model Selection
3.2
Model Training and Compilation
4
Application Development
4.1
System Flow
4.2
Object Tracker
4.3
Dashboard and Bounding Boxes Drawing
4.4
Physical Demo Setup
5
Performance Analysis
5.1
System Accuracy
5.2
Frame Rate
5.3
Cores Utilization
5.4
Power Consumption
6
Summary
7
References
3
Model Selection and Training