Quality

The quality node measures perceptual quality metrics of an image or video — resolution, blurriness (via a Laplacian measure) and bitrate — so low-quality assets can be flagged or filtered.

Kind

quality

Applies to

Image, Video

Inputs

None

Output keys

blurriness, image_width / image_height, video_width / video_height / video_fps / video_frame_count, quality_flag

Requirements

OpenCV / video4j native runtime. CPU-bound. No GPU required.

Persists to

asset_json_comp + asset_node_result ledger

Configuration

QualityNodeOptions toggle individual checks:

Option Meaning

checkBlurriness

Compute a blurriness score (Laplacian variance)

checkResolution

Record image/video dimensions

checkVideoBitrate

Measure the video bitrate

checkAudioBitrate

Measure the audio bitrate

Use Cases

  • Quality gating — pair with the quality filter to route low-quality assets differently or skip them.

  • Blurry-image detection — flag out-of-focus photos.

  • Library stats — resolution/bitrate reporting across an asset pool.