Asset Recognition in Content Collections

Asset recognition can process the following types of content from incoming posts:

• Images
  jpg, jpx, png, gif, webp, cr2, tif, bmp, jxr, psd, ico, heic

• Videos
  mp4, m4v, mkv, webm, mov, avi, wmv, mpg, flv
  When processing videos, the first minute of the video is processed at 60 FPS. After 1 minute, processing stops. This means that if the first minute of a video is modified, even if the rest of the content in the video is the same as an asset in your collection, there is a high likelihood the model will be unable to match the asset in the post with the appropriate label.

Collection level asset recognition is useful for tracking all of the assets you use in a campaign across your different social channels. When you activate asset recognition for a collection, the AI analyzes all of the assets in that collection. After the analysis completes, the following actions happen:

• Each asset is automatically assigned a shared label that identifies them as being part of that collection. The format of the label is predictable and easily recognizable:

  • NONE

    Format-
    [AssetRecognition]_[Collection:{name of collection}] {unique identifier}
    
    Example:
    [AssetRecognition]_[Collection:Streaming campaign] c1133700-123a-485c-8b55-476984f3139f

    After the label is automatically generated, you can rename the label in Settings → Labels.

• Whenever a public or private profile connected to the Emplifi platform publishes a post, that post is analyzed using the matching precision you previously specified when you turned on asset recognition for the collection.

• If the asset in an incoming post is within the matching precision setting you previously set, then the AI matches that asset with an asset in your collection and automatically assigns the identifying label to the post.

To create a unique label for a specific asset, you can also activate asset recognition at the level of the asset you want to track.

Asset recognition at the asset level is useful for tracking a specific asset across your different social channels. Unlike asset recognition at the asset level, when you activate asset recognition for an individual asset, the asset recognition model processes only that asset. After the analysis completes, the following actions happen:

• The asset is automatically assigned a unique label that identifies the asset. The format of the label is predictable and easily recognizable:

  • NONE

    Format-
    [AssetRecognition]_[Asset:{name of the asset}] {unique identifier}
    
    Example:
    [AssetRecognition]_[Asset:AdobeStock_488596189.jpg] 071a674d-858b-45ac-bdeb-bca028d83646

    After the label is automatically generated, you can rename the label in Settings → Labels.

• Whenever a public or private profile connected to the Emplifi platform publishes a post, that post is analyzed using the matching precision setting you previously specified when you turned on asset recognition for the asset.

• If the asset in an incoming post is within the matching precision setting you previously set, then the AI matches that asset with the asset in your collection and automatically assigns the identifying label to the post.

To instead create a label that is shared across all assets in the collection, you can also activate asset recognition at the level of the collection.

When you first activate asset recognition for either an asset or a collection of assets, you can set the level of precision the AI should use when matching posts assets to your collection assets. Before setting your threshold, consider the following:

• Asset recognition processes text as part of the image but is unable to understand language. This means that for the purpose of matching an incoming asset to an asset in your collection, the AI will match assets where the only difference is the text on the asset. For example, if you have an asset that uses a different language or call to action depending on the region the asset is posted, the AI will match these assets.

• The quality/resolution/cropping of an asset generally has no impact on the ability of the AI to match assets. For example, if someone shares a screenshot of a screenshot of an asset from your collection, the AI can still match the asset to the one in your collection despite the loss of quality.

• The priority when processing video is the frame data from the first minute of the video. The video thumbnail is used only as a last resort when the model is otherwise unsure if there is a match or not.

There are three options:

• High precision
  High precision detects identical assets or assets that are nearly identical (the assets only have slight alterations).

• Moderate precision
  Moderate precision detects identical assets, assets that are nearly identical, and assets that are mostly identical.

• Flexible precision
  Flexible precision detect identical assets, assets that are nearly identical, assets that are mostly identical, and assets that are similar but have a greater degree of variation.

Steps:

1. From the main navigation window, go to Content → Collections.
   The Collections list opens.

2. Click Create Collection.
   The New collection panel opens. For more information on creating collections, see Content Collections.

3. Click the Asset Recognition toggle to activate asset recognition for the collection.

4. Select the Matching precision:

   • High precision

   • Moderate precision

   • Flexible precision

5. Click Create Collection.
   A new, empty collection is created and opened.

As you add assets to the collection, they will be analyzed and labeled.

Steps:

1. From the main navigation window, go to Content → Collections.
   The Collections list opens.

2. Open an existing collection.

3. Click the Asset Recognition toggle to activate asset recognition for the collection.

4. Select the Matching precision:

   • High precision

   • Moderate precision

   • Flexible precision

After you set the matching precision, the AI will start labeling incoming posts with a collection level asset recognition label when it finds a match.

Steps:

1. From the main navigation window, go to Content → Collections.
   The Collections list opens.

2. Open an existing collection.

3. Click the asset you want to activate asset-level asset recognition for.

4. Click the Asset Recognition toggle to activate asset recognition for the asset.

5. Select the Matching precision:

   • High precision

   • Moderate precision

   • Flexible precision

After you set the matching precision, the AI will start labeling incoming posts with an asset level asset recognition label when it finds a match.

Steps:

1. From the main navigation window, go to Content → Collections.
   The Collections list opens.

2. Open your collection.

3. Determine which label you want to find:

   • To find posts that match the collection level asset recognition label, click Show Content in the Asset Recognition section of the General tab.
     The Content Feed opens and shows any posts that have the collection level asset recognition label.

   • To find posts that match the asset level asset recognition label, open an asset and click Show Content in the Asset Recognition section of the General tab.
     The Content Feed opens and shows any posts that have the asset level asset recognition label.
     

Steps:

1. From the main navigation window, go to Settings → Labels.
   The Labels list opens.

2. Click the name of the label.
   The label details panel opens.

3. In the This label was auto-generated by AI Asset Recognition in Collections banner at the top of the label details panel, click Open collection or Open asset.
   The collection or asset opens.

Steps:

1. From the main navigation window, go to Content → Collections.
   The Collections list opens.

2. Open your collection.

3. Determine which label you want to analyze:

   • To analyze posts that match the collection level asset recognition label, click Analyze in the Asset Recognition section of the General tab.
     The Campaigns & labels board of Unified Analytics opens and shows data for posts that have the collection level asset recognition label.

   • To analyze posts that match the asset level asset recognition label, open an asset and click Analyze in the Asset Recognition section of the General tab.
     The Campaigns & labels board of Unified Analytics opens and shows data for posts that have the asset level asset recognition label.

Alternatively, you can click the Analytics tab at either the collection level or asset level to receive data about posts with the label from within the collection itself.

Deleting an asset recognition label disables asset recognition for the collection or asset associated with the label. All analysis related to that label is also permanently erased.

Steps:

1. From the main navigation window, go to Settings → Labels.
   The Labels list opens.

2. Click the name of the label you want to delete.
   The label details panel opens.

3. Click the trashcan icon next to the label name.
   A warning popup opens.

4. Clicking Delete confirms you understand that deleting the label disables asset recognition and erases all data related to the label.
   The label is deleted. Asset recognition is disabled for the collection or asset associated with the label and all asset recognition data related to the label is also permanently erased.