Computer Vision on AWS by Lauren Mullennex & Nate Bachmeier & Jay Rao

Author:Lauren Mullennex & Nate Bachmeier & Jay Rao
Language: eng
Format: epub
Publisher: Packt
Published: 2023-11-15T00:00:00+00:00

Depending on where you import your images from, your images might not be labeled. You can use the Lookout for Vision console to add, change, and assign labels. You do need to label the dataset images according to the type of model you plan to use (classification versus segmentation). The labels determine the model type that Lookout for Vision will create.

There are a few other things to keep in mind while creating datasets for a project:

All images in a dataset must have the same dimension (otherwise, model training will fail).

Each of these images must contain only one type of anomaly. Providing images with multiple anomalies in a single image will reduce the model performance and increase the number of incorrect predictions.

Try to provide images with consistent conditions, such as the position of the camera (the same angle, zoom, quality, and so on), lighting, and pose of the object.

If you are providing separate datasets for training and testing, make sure to avoid duplicates.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.