Boundary detection evaluation
Illustrative embodiments are directed to a method and apparatus for evaluating boundary detection in an image. A processed image is received, wherein a detected boundary of an image of an object is identified in the processed image. A Radon transform is applied to the processed image for a plurality of angles to form a processed image histogram corresponding to the detected boundary for each of the plurality of angles. The processed image histogram for each of the plurality of angles and a corresponding ground truth histogram for each of the plurality of angles is normalized to provide a normalized processed image histogram and a normalized ground truth histogram for each of the plurality of angles, wherein the ground truth histogram corresponds to a ground truth boundary of the object for a corresponding angle. An indication of the edges of the normalized processed image histogram for each of the plurality of angles is plotted to form a boundary detection evaluation visualization.
1. A method of evaluating boundary detection in an image, comprising:
receiving a processed image, wherein a detected boundary of an object image of an object is identified in the processed image;
applying a Radon transform to the processed image for a plurality of angles to form a processed image histogram corresponding to the detected boundary for each of the plurality of angles;
normalizing the processed image histogram for each of the plurality of angles and a corresponding ground truth histogram for each of the plurality of angles to provide a normalized processed image histogram and a normalized ground truth histogram for each of the plurality of angles, wherein the ground truth histogram corresponds to a ground truth boundary of the object for a corresponding angle; and
plotting an indication of the edges of the normalized processed image histogram for each of the plurality of angles to form a boundary detection evaluation visualization,
wherein normalizing the processed image histogram and the ground truth histogram for each of the plurality of angles comprises normalizing the processed image histogram and the ground truth histogram such that edges of the normalized ground truth histogram for each of the plurality of angles are the same distance from an origin.
2. The method of claim 1 , further comprising displaying an indication of the ground truth boundary on the boundary detection evaluation visualization as a circle centered at the origin and having a radius equal to the distance.
3. The method of claim 2 , wherein normalizing the processed image histogram and the ground truth histogram for each of the plurality of angles comprises normalizing the processed image histogram and the ground truth histogram such that edges of the normalized ground truth histogram for each of the plurality of angles are at −1 and 1 with respect to the origin; and
displaying an indication of the ground truth boundary on the boundary detection evaluation visualization as a unit circle centered at the origin.
4. The method of claim 1 further comprising repeating the steps of receiving the processed image, applying the Radon transform, normalizing, and plotting for a plurality of different processed images that have been processed using a same boundary detection algorithm.
5. The method of claim 1 further comprising repeating the steps of receiving the processed image, applying the Radon transform, normalizing, and plotting for a same image processed using a plurality of different boundary detection algorithms.
6. An apparatus, comprising:
a receiver configured to receive a processed image, wherein a detected boundary of an image of an object is identified in the processed image;
a Radon transformer configured to apply a Radon transform to the processed image for a plurality of angles to form a processed image histogram corresponding to the detected boundary for each of the plurality of angles;
a normalizer configure to normalize the processed image histogram for each of the plurality of angles and a corresponding ground truth histogram for each of the plurality of angles to provide a normalized processed image histogram and a normalized ground truth histogram for each of the plurality of angles, wherein the ground truth histogram corresponds to a ground truth boundary of the object for a corresponding angle; and
a visualization generator configured to plot an indication of the edges of the normalized processed image histogram for each of the plurality of angles to form a boundary detection evaluation visualization,
wherein the normalizer is configured to normalize the processed image histogram and the ground truth histogram such that edges of the normalized ground truth histogram for each of the plurality of angles are the same distance from an origin.
7. The apparatus of claim 6 , wherein the visualization generator is configured to display an indication of the ground truth boundary on the boundary detection evaluation visualization as a circle centered at the origin and having a radius equal to the distance.
8. The apparatus of claim 7 , wherein:
the normalizer is configured to normalize the processed image histogram and the ground truth histogram such that edges of the normalized ground truth histogram for each of the plurality of angles are at −1 and 1 with respect to the origin; and
the visualization generator is configured to display an indication of the ground truth boundary on the boundary detection evaluation visualization as a unit circle centered at the origin.