Finalization of plots

Fixes #22 (closed).

Requires merge bob.measure!65 (merged).

Requires bob.measure!67 (merged)

marked as a Work In Progress

changed the description

added 3 commits

• 0f078e36 - [script][figure] Correct typo in criteria names
• d9af79ba - [script][vuln_commands] Typo in criteria
• 934d7680 - [test][test_commands] Fix test with new det impl

Compare with previous version

added 3 commits

• 620b00f9 - [script][figure] Refoctoring of DET and add ROC with same base
• 8cfacc26 - [setup] Add vuln roc command
• 7371a023 - [script][vuln_commands] Add ROC Command

Compare with previous version

unmarked as a Work In Progress

added 2 commits

• c5f8650a - [script][figure] Remove hardcoded markersize
• e4fc2242 - [doc][experiments] Update pad metrics table and rename test->eval in

Compare with previous version

Thanks @amohammadi for the review.

added 3 commits

• 939d2d7d - [doc][experiments] Correct Bp to BPCER
• 25a9c4d6 - [script][figure] Correct BP to BPCER
• 8136e03a - [doc][experiments] / [script][figure] Rename HTER tp ACER

Compare with previous version

resolved all discussions

• Please put bob pad epc in bob pad evaluate
• bob pad epc help message is not updated
• Only 4 column score formats exists in bob.pad.base. Please remove traces of 5 column.
• Pls add -v to all command execution examples.
• bob vuln commands show bob pad in their usage
• bob vuln metrics should show fnmr instead of bpcer
• bob pad epc: replace HTER with ACER
• bin/bob vuln epsc fake_scores/{licit,spoof}/scores-{dev,eval} -D takes forever to finish. Just run bin/bob vuln gen fake_scores (after bugfix, see below) to generate the scores.
• bin/bob vuln gen fake_scores neg and pos were swapped. I will push a fix for this
• bob vuln {roc,det} does not look like what @pkorshunov had.
• bob vuln epc does not seem correct. We have to investigate. IAPMR is always 100% and WER just goes up.
• bob vuln evaluate does not plot bob vuln roc
• Maybe we should rename these bob ... evaluate commands to bob ... eval. It's easier to write. The python functions can still be named evaluate but the exposed command name could be eval.
• bob vuln {roc,det} should produce separate figures for dev and eval by default. Right now it uses licit-dev and spoof-eval which does not make sense.

We need to majorly change bob vuln det,roc and implement bob vuln scatter but I think it's easier to explain in person than writing here.

added 1 commit

• 5222c861 - Fix score generation command

Compare with previous version

added 7 commits

• d0fb5821 - [script][figure] Changes names and defaults
• fe3ced43 - [script][pad_commands] Add -v on examples and remove 5-col ref
• 664a694f - [script][vuln_commands] Chanhe pad to vuln
• 849d0911 - [script][pad] Add prefix aliasing
• 371551f0 - [script][pad_commands] Add prefix aliasing
• 852d4da5 - [script][vuln] Add prefix aliasing and fix doc
• 22047462 - [script][vuln_commands] Add prefix aliasing and remove vuln command

Compare with previous version

added 1 commit

• 463f1d17 - [script][figure] Change vuln labels for det

Compare with previous version

added 2 commits

• 6c33b757 - [script][vuln_commands] Change defaults
• e1e9a249 - [script][figure] Fix det and roc plots

Compare with previous version

added 1 commit

• aed05f24 - [script][figure,pad_commands,vuln_commands] --titles option for PAD and

Compare with previous version