Zebra Technologies Corporation · NASDAQ:ZBRA
Zebra Technologies — Technical Presentation (2023)
22 slides from Zebra Technologies's technical presentation, 2023
Slide 1: Deep Learning OCR: Create a cross-platform algorithm in 10 minutes
Deep Learning OCR: Create a cross-platform algorithm in 10 minutes Jakub Cieplak Application Engineer 1
Slide 2: Aurora Vision™ Software products for machine vision
Aurora Vision™ Software products for machine vision ZEBRA TECHNOLOGIES
Slide 3: Product portfolio
Product portfolio Aurora Vision offers a comprehensive range of software products for image analysis Studio Library Deep Learning Graphical, dataflow-based environment For C++ and .NET Sample-based training for rapid development A set of ready-to-use functions for A set of seven ready-made tools bas
Slide 4: Value Proposition
Value Proposition Why Zebra Aurora Vision? • Rapid development – significantly faster compared to low-level programming • Easy-to-use, graphical programming environment • But still so powerful - over 3000 reliable, field-tested tools optimized for demanding applications • Mature solution – all tools
Slide 5: All-in-one development platform
All-in-one development platform Reduce your time to market by taking advantage of user-friendly and hardware-agnostic software Generate code or design user interface Develop an algorithm or train AI-based tools Select hardware ZEBRA TECHNOLOGIES
Slide 6: Multiple platforms supported
Multiple platforms supported Development environment: Windows Linux Aurora Vision Studio Aurora Vision Library *.avexe .Net code C++ code Execution enviroment: Windows Linux Android ZEBRA TECHNOLOGIES
Slide 7: Multiple technologies in one environment
Multiple technologies in one environment • Traditional 2D inspections 2D • Tools to analyse 3D data • Powerful deep learning algorithms 3D • All those technologies can be easily connected to work together Deep Learning ZEBRA TECHNOLOGIES
Slide 8: Different data types can be easily converted and used interchangeably
Different data types can be easily 3D converted and used interchangeably 2D 1D ZEBRA TECHNOLOGIES
Slide 9: Aurora Vision Studio
Aurora Vision Studio Overview – hands on ZEBRA TECHNOLOGIES
Slide 10: Deep Learning
Deep Learning ZEBRA TECHNOLOGIES
Slide 11: Deep Learning
Deep Learning Available Deep Learning tools: • Detect Features • Locate Points • Classify Object • Detect Anomalies • Segment Instances • Locate Objects • Locate Text • Read Characters ZEBRA TECHNOLOGIES
Slide 12: Detect Features
Detect Features Feature Detection Also known as supervised defect detection. The user carefully marks pixels corresponding to defects on the training images. The tool then learns to distinguish defects or other specific features by looking for their key characteristics automatically extracted with c
Slide 13: Locate Points
Locate Points Point Location Quick and easy object detection tool that looks for characteristic points on the input image. Key features: • Data labeling is easy (one click per object) • Detects a point, not a region • Relatively fast execution ZEBRA TECHNOLOGIES
Slide 14: Classify Object
Classify Object Object Classification Whole image classification is the most basic deep learning tool. The tool analyzes an entire image and provides a name for the most prominent object in the scene. Key features: • Simple and very easy to use • Separates up to 50 object classes in industrial appli
Slide 15: Detect Anomalies
Detect Anomalies Anomaly Detection In Anomaly Detection there is no specific definition of a defect—the tool is trained with Good samples and then looks for deviations of any kind. From the machine learning point of view it is a more difficult task as the neural network is not able to infer knowledg
Slide 16: Segment Instances
Segment Instances Instance Segmentation This tool simultaneously detects and classifies several objects on the input image. And it is not only a bounding rectangle of an object that is detected but a complete mask which then can be used for further processing. Key features: • Laborious labeling proc
Slide 17: Locate Objects
Locate Objects Object location This tool is used to locate and classify one or multiple objects within an image. The result of this technique is a list of rectangles bounding the predicted objects with corresponding class predictions and confidence scores. Key features: • Allows to train multiple cl
Slide 18: Locate Text
Locate Text Text location This tool is used to locate text on images. The result is a list of rectangles bounding text characters. Often used to detect character orientation before reading the actual writing. Key features: • Pre-trained, ready-to-use without any training • A good starting point for
Slide 19: Read Characters
Read Characters Character identification This tool is used to locate individual Complex backgrounds characters and detect appropriate symbols. Key features: Blurred text • Pre-trained, ready to use without any training • Extremely robust Reflective surfaces Low quality prints ZEBRA TECHNOLOGIES
Slide 20: Aurora Vision Studio: Deep learning OCR - hands on
Aurora Vision Studio Deep learning OCR - hands on ZEBRA TECHNOLOGIES
Slide 21: Questions
Questions ZEBRA TECHNOLOGIES
Slide 22: Thank You
Thank You ZEBRA and the stylized Zebra head are trademarks of Zebra Technologies Corp., registered in many jurisdictions worldwide. All other trademarks are the property of their respective owners. ©2023 Zebra Technologies Corp. and/or its affiliates. All rights reserved. ZEBRA TECHNOLOGIES