Ground Truth

Get started with Ground Truth

• From Unlabeled Data to a Deployed Machine Learning Model: A SageMaker Ground Truth Demonstration for Image Classification
• Run a Ground Truth labeling job
• Analyze Ground Truth labeling job results
• Compare Ground Truth results to known, pre-labeled data
• Train an image classifier using Ground Truth labels
• Deploy the Model
• Review

Task types

3D point clouds

• Create a 3D Point Cloud Labeling Job with Amazon SageMaker Ground Truth
• Create a 3D Point Cloud Labeling Job with Amazon SageMaker Ground Truth
• Acknowledgments

Annotation consolidation

• Understanding Annotation Consolidation: A SageMaker Ground Truth Demonstration for Image Classification

Object detection

• From Unlabeled Data to a Deployed Machine Learning Model: A SageMaker Ground Truth Demonstration for Object Detection
• Introduction
• Run a Ground Truth labeling job
• Analyze Ground Truth labeling job results
• Compare Ground Truth results to standard labels
• Train an object detection model using Ground Truth labels
• Deploy the Model
• Review

Pretrained model labeling

• Using a Pre-Trained Model for Cost Effective Data Labeling
• Introduction
• Iteration #1: Create Initial Labeling Job
• Iteration #2: Labeling Job with Pre-Trained Model
• Iteration #3: Second Data Subset Without Pre-Trained Model
• Conclusion

Bring your own model

• Part 1: Create Resources Needed for an Active Learning Workflow
• Part 2: Bring Your Own Model to an Active Learning Workflow
• Preprocessing and Tokenizing the data
• Adding the Containerized ML Model to ECR
• Training our Model
• Print the Image name in ECR

Analysis tasks

Data analysis of image classification

• Data Analysis Using a Ground Truth Image Classification Output Manifest

Worker labeling efficiency