About Lesson
This video describes the process of using Lineal Images to review and suppress non–relevant images.
Lineal Images is run as an Amplify job and classifies image files within a document based on their content.
The “Lineal Images Processing Status” field indicates a document has been identified as an image and assessed.
The categories associated with the image are reflected in the review workspace under the “Lineal Image Categories” field and can be used to narrow or prioritize the set for assessment.
The Lineal Images review interface can be launched via a mass operation on any population.
Images are grouped as tiles in each category and the categories are sorted from largest to smallest.
Specific categories can be selected via the search bar. Embedded logos and graphics are often among the highest volume of images in a population and are grouped.
Lineal Images will identify if an image is duplicative of others and only show that image once while indicating the number of duplicates. Any tagging applied to that image will be applied to all duplicates.
Lineal Images also reflects how certain the AI is that the image falls within the defined category and what other categories that image is within.
Once a category with non–relevant images for suppression is identified those images can be used to determine other categories that should be prioritized.
Tags within the workspace can be added to the interface and tagging is
automatically reflected in the same field in the review environment.
automatically reflected in the same field in the review environment.
Individual images can be selected to review, or multiple images can be chosen for mass tagging.
Categories can be scanned from left to right and additional images added to the view by clicking the arrow button.
The number of images visible in a category is reflected next to a check box and can be selected for mass-tagging.
An entire category can also be tagged at once. A pop–up indicates if the action includes instances not yet visible in the category view.