search

marialeyvallina / generalized_contrastive_loss

MIT License
87 stars 14 forks source link

How to get the value of Extended CMU Seasons? #15

Closed MAX-OTW closed 1 year ago

MAX-OTW commented 1 year ago

Hello, @marialeyvallina @nicstrisc ，I really appreciate for your work. I have a question, the Table 6 shows that the result of Extended CMU Seasons of ResNext-GeM-GCL: Urban:11.1 / 28.7 / 87.4; Suburban:4.2 / 14.6 / 77.4; Park:3.1 / 11.1 / 57.7

And the Mean result is : 6.1 / 18.2 / 74.9

Is it weighted in some way? How is it converted? I am always looking forward to your kind response. Best regards.

MAX-OTW commented 1 year ago

By the way, how do you convert the results of RobotCar Seasons v2? Could you please tell me the specific weighted average of these two datasets(RobotCar Seasons v2 and Extended CMU Seasons)? I am always looking forward to your kind response. Best regards.

marialeyvallina commented 1 year ago

Hi. For both CMU and Oxford it is weighted by the amount of images in each subset. As you can see from the dataset Oxford has 1443 query images in day and 429 in night. The avg is (day_result1443+night_result429)/(1443+429). With Extendedcmu is the same but for the amount of images in their respective subsets.