ubc-vision/image-matching-benchmark

Mistake in text description of the pose error metric

Closed this issue · 1 comments

The sentence

To measure the accuracy for a given pair of images, we compute the angular error between the estimated ground truth translation vectors and the estimated and ground truth translation vectors, and take the largest of the two.

doesn't seem quite right, does it?

The main metric is the quality of the estimated poses. We measure them in relative terms, as stereo poses can be recovered up to a scale factor. To measure the accuracy for a given pair of images, we compute the angular error between the estimated ground truth translation vectors and the estimated and ground truth translation vectors, and take the largest of the two. We threshold this error to reach a binary decision, e.g. is this estimate accurate at 5 or 10 degrees, and accumulate it over multiple image pairs to compute the "average accuracy". We do this for multiple error thresholds, from 1 to 10 degrees with a 1-degree resolution. Finally, we compute the area under this curve: this gives us a single scalar value, that we call "mean Average Accuracy" or mAA, analogously to the mean Average Precision or mAP commonly used in object detection.

Thanks, fixed. One of the translation should be rotation