Implement more metrics

[pesekon2]

Research on what metrics should be used to evaluate the chosen architectures should be done.

Proposed metrics:

• Accuracy (see [1], [4])
• Precision (see [1], [2], [3])
• Recall (see [1], [2], [3])
• IoU (see [1])
• mIoU
• Possible less common (maybe):
  • F-score (see [1], maybe [3])
  • F1-score (see [1], [2])
  • kappa (see [3], [4])
  • edge overall accuracy (see [5])
  • edge omission error (see [5])
  • edge comission error (see [5])

[1] https://apps.webofknowledge.com/full_record.do?product=WOS&search_mode=GeneralSearch&qid=5&SID=C2sVCKfXpvWdERfQXan&page=1&doc=2&cacheurlFromRightClick=no
[2] https://apps.webofknowledge.com/full_record.do?product=WOS&search_mode=GeneralSearch&qid=27&SID=F62Uul2VCwKAvqdqj7a&page=1&doc=1&cacheurlFromRightClick=no
[3] https://apps.webofknowledge.com/full_record.do?product=WOS&search_mode=GeneralSearch&qid=7&SID=D4bn73FMEr9wrkiA1vO&page=1&doc=5&cacheurlFromRightClick=no
[4] https://www.scopus.com/record/display.uri?eid=2-s2.0-85047519048&origin=resultslist&sort=cp-f&src=s&st1=%22remote+sensing%22+AND+%22convolutional+neural+network%22&nlo=&nlr=&nls=&sid=bfaa9680c87071569ac0d16ba3a0ddbc&sot=b&sdt=cl&cluster=scoopenaccess%2c%221%22%2ct%2bscopubyr%2c%222020%22%2ct%2c%222019%22%2ct%2c%222018%22%2ct%2bscosubtype%2c%22re%22%2ct&sl=56&s=ALL%28%22remote+sensing%22+AND+%22convolutional+neural+network%22%29&relpos=15&citeCnt=45&searchTerm=
[5] https://apps.webofknowledge.com/full_record.do?product=WOS&search_mode=GeneralSearch&qid=2&SID=E179BseTTc62l1cT1uZ&page=1&doc=1