关于分割的class

[lyh19951201]

你好，我想请问您，为什么在分割的时候是9类，而不是15类，label里面最大不是14嘛，还有就是如果是9类那么比9大的数应该怎么处理呢？希望刘博士能给解答，万分感谢！！

[wangqh1996]

您好，我也想问同样的问题，就是现在给出的分割类别应该是MFNet数据集的吧，针对FMB数据集的这些类别的PALETTE和CLASS具体是多少，可以给出吗？

[LiuZhu-CV]

是MFNet数据集的

[LiuZhu-CV]

The class and related plates are defined as

dict = {
0: (0, 0, 0),
1: (179, 228, 228), # road
2: (181, 57, 133), # sidewalk
3: (67, 162, 177), # building
4: (200, 178, 50), # lamp
5: (132, 45, 199), # sign
6: (66, 172, 84), # vegetation
7: (179, 73, 79), # sky
8: (76, 99, 166), # person
9: (66, 121, 253), # car
10: (137, 165, 91), # truck
11: (155, 97, 152), # bus
12: (105, 153, 140), # motocycle
13: (222, 215, 158), # bicycle
14: (135, 113, 90), # pole
}

[LiuZhu-CV]

你好，我想请问您，为什么在分割的时候是9类，而不是15类，label里面最大不是14嘛，还有就是如果是9类那么比9大的数应该怎么处理呢？希望刘博士能给解答，万分感谢！！

config 文件里对应修改就行了

[LiuZhu-CV]

The plate of FMB in the paper is defined as

palette = np.array([(0, 0, 0), (173, 229, 229), (187, 57, 134), (45, 163, 178), (206, 176, 47), (131, 54, 200), (56, 171, 83), (183, 71, 78), (66, 102, 167), (14, 127, 255), (138, 163, 91), (156, 98, 153), (101, 153, 140), (225, 214, 155), (136, 111, 89)])

[448357739]

The class and related plates are defined as

dict = {
0: (0, 0, 0),
1: (179, 228, 228), # road
2: (181, 57, 133), # sidewalk
3: (67, 162, 177), # building
4: (200, 178, 50), # lamp
5: (132, 45, 199), # sign
6: (66, 172, 84), # vegetation
7: (179, 73, 79), # sky
8: (76, 99, 166), # person
9: (66, 121, 253), # car
10: (137, 165, 91), # truck
11: (155, 97, 152), # bus
12: (105, 153, 140), # motocycle
13: (222, 215, 158), # bicycle
14: (135, 113, 90), # pole
}

您好，数据集label中似乎没有第13类bicycle的标签，是这类本来就不存在还是有别的label值而非13

[shenzw21]

您好，我也是这个问题，数据集label中似乎没有第13类bicycle的标签

[tianzhiya]

在哪修改dict = { 0: (0, 0, 0), 1: (179, 228, 228), # road 2: (181, 57, 133), # sidewalk 3: (67, 162, 177), # building 4: (200, 178, 50), # lamp 5: (132, 45, 199), # sign 6: (66, 172, 84), # vegetation 7: (179, 73, 79), # sky 8: (76, 99, 166), # person 9: (66, 121, 253), # car 10: (137, 165, 91), # truck 11: (155, 97, 152), # bus 12: (105, 153, 140), # motocycle 13: (222, 215, 158), # bicycle 14: (135, 113, 90), # pole } ？

[lxproot]

在哪修改dict = { 0: (0, 0, 0), 1: (179, 228, 228), # road 2: (181, 57, 133), # sidewalk 3: (67, 162, 177), # building 4: (200, 178, 50), # lamp 5: (132, 45, 199), # sign 6: (66, 172, 84), # vegetation 7: (179, 73, 79), # sky 8: (76, 99, 166), # person 9: (66, 121, 253), # car 10: (137, 165, 91), # truck 11: (155, 97, 152), # bus 12: (105, 153, 140), # motocycle 13: (222, 215, 158), # bicycle 14: (135, 113, 90), # pole } ？

我也没有找到这个修改，请问你解决了吗

[LiuZhu-CV]

在哪修改dict = { 0: (0, 0, 0), 1: (179, 228, 228), # road 2: (181, 57, 133), # sidewalk 3: (67, 162, 177), # building 4: (200, 178, 50), # lamp 5: (132, 45, 199), # sign 6: (66, 172, 84), # vegetation 7: (179, 73, 79), # sky 8: (76, 99, 166), # person 9: (66, 121, 253), # car 10: (137, 165, 91), # truck 11: (155, 97, 152), # bus 12: (105, 153, 140), # motocycle 13: (222, 215, 158), # bicycle 14: (135, 113, 90), # pole } ？

我也没有找到这个修改，请问你解决了吗

在util/util.py修改

[lxproot]

在哪修改dict = { 0: (0, 0, 0), 1: (179, 228, 228), # road 2: (181, 57, 133), # sidewalk 3: (67, 162, 177), # building 4: (200, 178, 50), # lamp 5: (132, 45, 199), # sign 6: (66, 172, 84), # vegetation 7: (179, 73, 79), # sky 8: (76, 99, 166), # person 9: (66, 121, 253), # car 10: (137, 165, 91), # truck 11: (155, 97, 152), # bus 12: (105, 153, 140), # motocycle 13: (222, 215, 158), # bicycle 14: (135, 113, 90), # pole } ？

我也没有找到这个修改，请问你解决了吗

在util/util.py修改

万分感谢！