The Learn and Recognize actions support the generation of benchmark output. You enable this mode by passing the -b
switch to panoptes.
Below is an invocation of panoptes instructing it to print out benchmark data into a tab-delimited file for all recognized persons visible in all JPEG images in a specified directory:
> panoptes recognize <path to image(s)> -b [-Ckey=value ...]
The format of the benchmark output is:
<image file name> <person ID> <confidence percentage> <recognition time>
where:
Note: Confidence percentage can range from 0 to approximately 180. A value of 100 indicates near certainty of a match.
You can redirect the benchmark output to a file as shown below to open it in a spreadsheet application. The data is stored as a tab-separated list.
> panoptes recognize directory_with_jpegs -b > my_data.tsv