How we enhance AWS SageMaker Object Detection with “Mandarins”

Business Case

Data Preparation

Images of cars from various angles
Cyclops Tech

Building the Model

Jupyter Notebook
Monitoring validation and training accuracy via CloudWatch

Testing the Model

False Positive error rate at various confidence score thresholds
Odometer counter was mistakenly identified as a rego plate with a high confidence score of 0.985
Positive sample (left), Negative sample (right)

Solution

Random size mandarin is placed at random location in images when there is no rego
False positive error rate comparison between models at various confidence score threshold
Precision and Recall curve comparison between models at various confidence score thresholds

Summary

Credits

--

--

--

The AI guy Connect with me on https://www.linkedin.com/in/agustinus-nalwan

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

How do Netflix recommendations work?

Using Machine Learning in Semiconductor Industry to Predict the Yield in R

Send image notification on person detection — using OpenCV and Python

ML Best Practices: Test Driven Development at Latent Space

The top 10 ML algorithms for data science in 5 minutes

How I end up using Machine Learning in Optics/Photonics Applications for my PhD

Question and Answering vs Machine Reading Comprehension (QA vs MRC)

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Agustinus Nalwan

Agustinus Nalwan

The AI guy Connect with me on https://www.linkedin.com/in/agustinus-nalwan

More from Medium

Train a Custom Computer Vision Model Using OCI Vision Service

Spam Detection: Train in one language, Predict in another language? | Zero-Shot Learning | Pytorch

Skew correction in Documents using Deep learning.