Filter options

Filtering allows you to classify and sort images to display a subset of images. Filtering is useful when there are many images and you want to process a specific subset of image at a time. Filtering offers multiple options that can be used together and separately to display only a subset of images.

You can filter the images based on the following options:

Label

• All

• Alarms – This option displays for an Anomaly Detection skill.

• Good Data – This option displays for an Anomaly Detection skill.

• Label 1 – This option displays for a Discrete State Detection skill. For example, Good.

• Label 2 – This option displays for a Discrete State Detection skill. For example, Bad.

Feedback

• All

• No Feedback – This option when selected only displays images that have not been reclassified by the user.

• Has Feedback – This option when selected only displays images that have been reclassified by the user.

Time

• All

• Before – Use the calendar to provide a date. This option displays all predictions before the selected date.

• After – Use the calendar to provide a date. This option displays all predictions after the selected date.

• Between – Use the calendar to provide two dates. This option displays all predictions between the selected dates.

Sort

• Date Ascending – The option displays the predictions sorted from the earliest date available.

• Date Descending – The option displays the predictions sorted from the most recent date available. For example, the current date.

• Score Ascending – This option displays the predictions sorted from the lowest confidence score.

• Score Descending – This option displays the predictions sorted from the highest confidence score.

To filter image predictions

1. Select (Filter).

   The Filter side sheet appears.

2. Select an option from at least one box.

   The Review page only displays the images that match the criteria.

   (Filter) displays a badge with the number of images that match the filter criteria.

3. Select Reset to clear the selected option.