lim-anggun / FgSegNet_v2

FgSegNet_v2: "Learning Multi-scale Features for Foreground Segmentation.” by Long Ang LIM and Hacer YALIM KELES
https://arxiv.org/abs/1808.01477
Other
149 stars 43 forks source link

the accuracy of a catogery is very low the port_0_17fps : 0.435026037734113 #16

Open slowlyideal opened 4 years ago

slowlyideal commented 4 years ago

it is very strange,the most are 99% and 98%,a liittle 96% but the port_0_17fps : 0.435026037734113

Akhilesh64 commented 4 years ago

@slowlyideal same problem here. Did you tweak anything or are using everything default ? The training set seems to be the problem to me here, there arent many features to learn. Using a threshold of 0.9 will yield an f-score of about 0.54 without training frames with training frames included its around 0.79

@lim-anggun can you share your stats file.