opencv assertion_error ai_generated true

cv::error: (-215:Assertion failed) !nms_iou_threshold.empty() && nms_iou_threshold > 0 && nms_iou_threshold <= 1 in function 'NMSBoxes'

ID: opencv/dnn-softnms-parameter-error

Also available as: JSON · Markdown · 中文
88%Fix Rate
85%Confidence
1Evidence
2023-03-15First Seen

Version Compatibility

VersionStatusIntroducedDeprecatedNotes
4.5.5 active — — —
4.6.0 active — — —
4.7.0 active — — —
4.8.1 active — — —
4.9.0 active — — —

Root Cause

Non-maximum suppression (NMS) called with an invalid or out-of-range IoU threshold parameter.

generic

中文

非极大值抑制 (NMS) 调用时使用了无效或超出范围的 IoU 阈值参数。

Official Documentation

https://docs.opencv.org/4.x/d6/d0f/group__dnn.html

Workarounds

  1. 95% success Ensure nms_iou_threshold is a float between 0.01 and 1.0, e.g., cv::NMSBoxes(boxes, scores, 0.5, 0.4) where 0.4 is the IoU threshold. 
    Ensure nms_iou_threshold is a float between 0.01 and 1.0, e.g., cv::NMSBoxes(boxes, scores, 0.5, 0.4) where 0.4 is the IoU threshold.
  2. 90% success Validate the parameter before calling NMSBoxes: if (iou_threshold <= 0.0 || iou_threshold > 1.0) { iou_threshold = 0.5; } 
    Validate the parameter before calling NMSBoxes: if (iou_threshold <= 0.0 || iou_threshold > 1.0) { iou_threshold = 0.5; }

中文步骤

  1. Ensure nms_iou_threshold is a float between 0.01 and 1.0, e.g., cv::NMSBoxes(boxes, scores, 0.5, 0.4) where 0.4 is the IoU threshold.
  2. Validate the parameter before calling NMSBoxes: if (iou_threshold <= 0.0 || iou_threshold > 1.0) { iou_threshold = 0.5; }

Dead Ends

Common approaches that don't work:

  1. Setting nms_iou_threshold to 0 to disable NMS 95% fail

    IoU threshold must be strictly greater than 0; 0 triggers the assertion.

  2. Using a negative value like -0.5 for aggressive suppression 100% fail

    Negative values fail the '> 0' check in the assertion.

  3. Passing a string or list instead of a float 80% fail

    The empty() check on a non-scalar type passes, but the comparison fails with type error.