​
What Are blurry Images and Objects?

Blurry images lack sharpness and clarity, often making it difficult to recognize objects or extract visual detail. This can significantly reduce the performance of machine learning models, lower data quality, and hinder accurate human annotation.

​
Common Causes of Blurriness

  • Focus issues: Autofocus or manual focus failures can result in out-of-focus shots.
  • Motion blur: Movement of the subject or camera during exposure leads to streaking or distortion.
  • Camera shake: Handheld capture or slow shutter speeds often introduce unintended blur.
  • Compression artifacts: Over-compressed images may lose edge detail and overall clarity.

​
Why It Matters

ProblemImpact
Visual interpretation errorsAnnotators may struggle to label blurry data, increasing labeling noise.
Model accuracy dropObject detection, OCR, and segmentation models perform poorly on blurred inputs.
Poor user experienceLow-quality images degrade trust and engagement in user-facing applications.

​
How to Detect and Remove Blurry Images

  • Detect Blurry Images:
    Go to “Add Filter” → “Quality Issue” → select “Blurry” → choose “IS” as the logic operator → set the desired confidence threshold (default is 0.5).
    Export the blurry images using Export → “Matching the applied filter.”
  • Remove Blurry Images:
    Go to “Add Filter” → “Quality Issue” → select “Blurry” → choose “IS NOT” as the logic operator → set the desired confidence threshold (default is 0.5).
    Export the images without blurry issues using Export → “Matching the applied filter.”
Managing blurriness helps improve data clarity, annotation quality, and downstream model performance—especially in high-precision environments.