Filter and Sort

Aligned Image with Page Break

Filtering, sorting, and searching data in Encord Active is crucial for various reasons. It enables insights and actionable results on the following key aspects and more:

  • Identification of patterns, trends, or anomalies within a subset of the data.

  • Recognition of duplicates, outliers, and inconsistencies.

  • Removal of irrelevant, noisy, and erroneous data.

  • Understanding a model's behaviour and potential skewness when facing different subsets of the data.

  • After filtering, sorting, and searching your data, use Collections with Annotate Projects to streamline your annotation process.

Encord Active provides the following methods:

  • Filter: Refine your searches using quality metrics, Collections, data types, annotation types, annotation classes, and by annotator from Annotate.

  • Sort: Sort your data/labels/predictions, in ascending or descending order, using quality metrics.

  • Natural language and image search: Enter descriptive queries in everyday language or use images to make it easier to find relevant images without the need for specific keywords or complex search parameters.

  • Embedding plot: A two-dimensional visualization technique used by Encord to represent high-dimensional data in a more interpretable form. Use the plot to select points within a specific rectangular area, thereby focusing on a particular subset of data points for in-depth analysis.

Filters

In the Active Explorer for a Project you can refine searches by data quality metrics, label quality metrics, Collections, data types, annotation types, annotation classes, and by annotator from Annotate.

👍

Tips

Filters provide the ability to include or exclude images based on your filtering criteria.

Data Quality Metrics

For more detailed information on Data Quality Metrics go here.

TitleMetric TypeOntology Type
Area - Ranks images by their area (width/height).image
Aspect Ratio- Ranks images by their aspect ratio (width/height).image
Blue Value - Ranks images by how blue the average value of the image is.image
Brightness - Ranks images by their brightness.image
Contrast- Ranks images by their contrast.image
Diversity - Forms clusters based on the ontology and ranks images from easy samples to annotate to hard samples to annotate.image
Frame Number - Selects images based on a specified range.image
Green Value- Ranks images by how green the average value of the image is.image
Height - Ranks images by the height of the image.image
Object Count - Counts number of objects in the image.imagebounding box, checklist, point, polygon, polyline, radio, rotatable bounding box, skeleton, text
Object Density - Computes the percentage of image area that is occupied by objects.imagebounding box, polygon, rotatable bounding box
Randomize Images - Assigns a random value between 0 and 1 to images.image
Red Value - Ranks images by how red the average value of the image is.image
Sharpness - Ranks images by their sharpness.image
Uniqueness - Finds duplicate and near-duplicate images.image
Width - Ranks images by the width of the image.image
Label Quality Metrics

For more detailed information on Label Quality Metrics go here.

TitleMetric TypeOntology Type
Absolute Area - Computes object size in amount of pixels.imagebounding box, polygon, rotatable bounding box
Aspect Ratio - Computes aspect ratios of objects.imagebounding box, polygon, rotatable bounding box
Blue Value - Ranks annotated objects by how blue the average value of the object is.imagebounding box, polygon, rotatable bounding box
Brightness - Ranks annotated objects by their brightness.imagebounding box, polygon, rotatable bounding box
Border Proximity - Ranks annotations by how close they are to image borders.imagebounding box, point, polygon, polyline, rotatable bounding box, skeleton
Broken Object Tracks - Identifies broken object tracks based on object overlaps.sequence, videobounding box, polygon, rotatable bounding box
Brightness - Ranks annotated objects by their brightness.imagebounding box, polygon, rotatable bounding box
Confidence - The confidence that an object was annotated correctly. While arguably not making much sense when annotated by a human, this value is very important for objects that were automatically labeled.imagebounding box, polygon, rotatable bounding box
Contrast - Ranks annotated objects by their contrast.imagebounding box, polygon, rotatable bounding box
Classification Quality - Compares image classifications against similar images.imageradio
Green Value - Ranks annotated objects by how green the average value of the object is.imagebounding box, polygon, rotatable bounding box
Height - Ranks annotated objects by the height of the object.imagebounding box, polygon, rotatable bounding box
Inconsistent Object Class - Looks for overlapping objects with different classes (across frames).sequence, videobounding box, polygon, rotatable bounding box
Inconsistent Track ID - Looks for overlapping objects with different track-ids (across frames).sequence, videobounding box, polygon, rotatable bounding box
Label Duplicates - Ranks labels by how likely they are to represent the same object.imagebounding box, polygon, rotatable bounding box
Missing Objects - Identifies missing objects based on object overlaps.sequence, videobounding box, polygon, rotatable bounding box
Object Classification Quality - Compares object annotations against similar image crops.imagebounding box, polygon, rotatable bounding box
Occlusion Risk - Tracks objects and detect outliers in videos.sequence, videobounding box, rotatable bounding box
Polygon Shape Anomaly - Calculates potential outliers by polygon shape.imagepolygon
Randomize Objects - Assigns a random value between 0 and 1 to objects.imagebounding box, polygon, rotatable bounding box
Red Value - Ranks annotated objects by how red the average value of the object is.imagebounding box, polygon, rotatable bounding box
Relative Area - Computes object size as a percentage of total image size.imagebounding box, polygon, rotatable bounding box
Sharpness - Ranks annotated objects by their sharpness.imagebounding box, polygon, rotatable bounding box
Width - Ranks annotated objects by the width of the object.imagebounding box, polygon, rotatable bounding box

Collections: Collections are a new way to save interesting groups of data units and labels, to support and guide your downstream workflow.

Custom Metadata: Custom metadata added to Annotate Projects. When an Annotate Project imports to Active, if that Annotate Project has custom metadata, the custom metadata is available to filter your data in Active.

👍

Tips

For information on importing custom metadata to an Annotate project, refer to Adding Metadata and Reading and Writing Metadata in the documentation.

Data Types: Data, labels, and Predictions can be filtered by: images, image sequences, image groups, and videos

Dataset: Datasets included in an Annotate Project.

Data Units: Data units (images, image groups, videos) included in a dataset.

Annotation Type:: Data, labels, and Predictions can be filtered by: classification, bounding box, rotatable bounding box, point, polyline, polygon, skeleton, and bitmask.

Class: Data, labels, and Predictions can be filtered by annotation class.

Annotator:: Data, labels, and Predictions can be filtered by the person who annotated the images/videos.


To filter data, labels, or predictions:

  1. Log in to the Encord platform.
    The landing page for the Encord platform appears.

  2. Click Active in the main menu.
    The landing page for Active appears.

  3. Click the Project.
    The landing page for the Project appears with the Explorer tab selected.
    Filter with example

  4. Select Data, Labels, or Predictions.

  5. Click Filters.
    The Filters tab appears.

  6. Click Add Filter.
    A menu appears.

  7. Add and configure the filters you need.
    Images/video filter in the Explorer workspace.

👍

Tips

  • Sort and use the NLP or image searches to further help get the results you want.

  • After filtering, sorting, and searching, create a Collection.

Preset Filters

Preset filters provide a way to save your filtering criteria for use and reuse on other Projects. Preset filters are made up of global and local filter criteria.

Global filter criteria are filters, and their settings, that can apply to any Project. For example, Data Quality Metrics like Area, Blue Value, and Sharpness or Label Quality Metrics like Annotation Quality, Confidence, or Label Duplicates.

Local filter criteria are filters that can only be applied to a specific Project. For example, the following filters and their settings are likely only applicable to a specific Project: Class, Dataset, or Collection.

Create a Preset Filter

Create Presets (Preset Filters and their settings) from the Filter tab in your Projects.

To create a Preset (Preset Filter and their settings):
  1. Log in to the Encord platform.
    The landing page for the Encord platform appears.

  2. Click Active in the main menu.
    The landing page for Active appears.

  3. Click the Project.
    The landing page for the Project appears with the Explorer tab selected.
    Filter with example

  4. Select Data, Labels, or Predictions.

  5. Click Filters.
    The Filters tab appears.

  6. Click Add Filter.
    A menu appears.

  7. Add and configure the filters you need.
    Images/video filter in the Explorer workspace.

  8. Click Presets > Create Presets once you have added all the filters you need and specified each filter's settings.
    After creating the Preset you can use the Preset in this or any other Project.

Use a Preset Filter in a Project

Once you have created one or more Preset Filters, you can apply them to any of your Projects from the Filter tab.

ℹ️

Note

Global filters apply to any Project, but Local filters only apply on the Project where the Preset was created.

To create a Preset (Preset Filter and their settings):
  1. Log in to the Encord platform.
    The landing page for the Encord platform appears.

  2. Click Active in the main menu.
    The landing page for Active appears.

  3. Click the Project.
    The landing page for the Project appears with the Explorer tab selected.
    Filter with example

  4. Select Data, Labels, or Predictions.

  5. Click Filters.
    The Filters tab appears.

  6. Click Presets.
    A menu appears.

  7. Click Load Presets.
    Presets appears on the Filter tab with a dropdown menu.
    Filter with example

  8. Select the Preset you want to use from the dropdown.
    Images/video frames filter in the Explorer workspace based on the Preset Filters and their settings.

ℹ️

Note

Global filters apply to any Project, but Local filters only apply on the Project where the Preset was created.

Sorting

Sort your data, in ascending or descending order, using data quality metrics.
Sort your labels or predictions, in ascending or descending order, using label quality metrics.

Data Quality Metrics

For more detailed information on Data Quality Metrics go here.

TitleMetric TypeOntology Type
Area - Ranks images by their area (width/height).image
Aspect Ratio - Ranks images by their aspect ratio (width/height).image
Blue Value - Ranks images by how blue the average value of the image is.image
Brightness - Ranks images by their brightness.image
Contrast - Ranks images by their contrast.image
Diversity - Forms clusters based on the ontology and ranks images from easy samples to hard samples.image
Green Value - Ranks images by how green the average value of the image is.image
Height - Ranks images by the height of the image.image
Object Count - Counts number of objects in the image.imagebounding box, checklist, point, polygon, polyline, radio, rotatable bounding box, skeleton, text
Object Density - Computes the percentage of image area that is occupied by objects.imagebounding box, polygon, rotatable bounding box
Randomize Images - Assigns a random value between 0 and 1 to images.image
Red Value - Ranks images by how red the average value of the image is.image
Sharpness - Ranks images by their sharpness.image
Uniqueness - Finds duplicate and near-duplicate images.image
Width - Ranks images by the width of the image.image
Label Quality Metrics

For more detailed information on Label Quality Metrics go here.

TitleMetric TypeOntology Type
Absolute Area - Computes object size in amount of pixels.imagebounding box, polygon, rotatable bounding box
Aspect Ratio - Computes aspect ratios of objects.imagebounding box, polygon, rotatable bounding box
Blue Value - Ranks annotated objects by how blue the average value of the object is.imagebounding box, polygon, rotatable bounding box
Brightness - Ranks annotated objects by their brightness.imagebounding box, polygon, rotatable bounding box
Border Proximity - Ranks annotations by how close they are to image borders.imagebounding box, point, polygon, polyline, rotatable bounding box, skeleton
Broken Object Tracks - Identifies broken object tracks based on object overlaps.sequence, videobounding box, polygon, rotatable bounding box
Brightness - Ranks annotated objects by their brightness.imagebounding box, polygon, rotatable bounding box
Confidence - The confidence that an object was annotated correctly. While arguably not making much sense when annotated by a human, this value is very important for objects that were automatically labeled.imagebounding box, polygon, rotatable bounding box
Contrast - Ranks annotated objects by their contrast.imagebounding box, polygon, rotatable bounding box
Classification Quality - Compares image classifications against similar images.imageradio
Green Value - Ranks annotated objects by how green the average value of the object is.imagebounding box, polygon, rotatable bounding box
Height - Ranks annotated objects by the height of the object.imagebounding box, polygon, rotatable bounding box
Inconsistent Object Class - Looks for overlapping objects with different classes (across frames).sequence, videobounding box, polygon, rotatable bounding box
Inconsistent Track ID - Looks for overlapping objects with different track-ids (across frames).sequence, videobounding box, polygon, rotatable bounding box
Label Duplicates - Ranks labels by how likely they are to represent the same object.imagebounding box, polygon, rotatable bounding box
Missing Objects - Identifies missing objects based on object overlaps.sequence, videobounding box, polygon, rotatable bounding box
Object Classification Quality - Compares object annotations against similar image crops.imagebounding box, polygon, rotatable bounding box
Occlusion Risk - Tracks objects and detect outliers in videos.sequence, videobounding box, rotatable bounding box
Polygon Shape Anomaly - Calculates potential outliers by polygon shape.imagepolygon
Randomize Objects - Assigns a random value between 0 and 1 to objects.imagebounding box, polygon, rotatable bounding box
Red Value - Ranks annotated objects by how red the average value of the object is.imagebounding box, polygon, rotatable bounding box
Relative Area - Computes object size as a percentage of total image size.imagebounding box, polygon, rotatable bounding box
Sharpness - Ranks annotated objects by their sharpness.imagebounding box, polygon, rotatable bounding box
Width - Ranks annotated objects by the width of the object.imagebounding box, polygon, rotatable bounding box

To sort data, labels, or predictions:

  1. Log in to the Encord platform.
    The landing page for the Encord platform appears.

  2. Click Active in the main menu.
    The landing page for Active appears.

  3. Click the Project.
    The landing page for the Project appears with the Explorer tab selected.
    Filter with example

  4. Select Data, Labels, or Predictions.

  5. Select the metric to sort the data.

  6. Specify ascending or descending order.

👍

Tips

  • Filter and use the NLP or image searches to further help get the results you want.

  • After filtering, sorting, and searching, create a Collection.


What’s Next