自己的单个数据是二维的及类别可以使用few shot 分类吗？ about libfewshot HOT 3 CLOSED

rl-vig commented on August 30, 2024

自己的单个数据是二维的及类别可以使用few shot 分类吗？

from libfewshot.

Comments (3)

wZuck commented on August 30, 2024

你是指数据的格式为 [C,H,W] = [1,H,W] 吗，如果是这样的话，可以。但是因为代码中都是对3通道图像做处理的，因此需要对dataloader或者backbone稍作修改。

from libfewshot.

henbucuoshanghai commented on August 30, 2024

感谢回复自己的单个数据是 1，h,w的维度   及类别   我把1维度扩展为三维，就可以和代码对齐了？感谢

from libfewshot.

wZuck commented on August 30, 2024

是的，你可以在

Lines 177 to 179 in f002bf3

 image_name = self.data_list[idx] 

 image_path = os.path.join(self.data_root, "images", image_name) 

 data = self.loader(image_path)

里修改图片载入逻辑，读取灰度图像并转换为三通道图像。

Lines 12 to 13 in f002bf3

 MEAN = [120.39586422 / 255.0, 115.59361427 / 255.0, 104.54012653 / 255.0] 

 STD = [70.68188272 / 255.0, 68.27635443 / 255.0, 72.54505529 / 255.0]

这里的MEAN和STD是在ImageNet上统计得到的参数，如有需要可以自行修改。

from libfewshot.

	image_name = self.data_list[idx]
	image_path = os.path.join(self.data_root, "images", image_name)
	data = self.loader(image_path)

	MEAN = [120.39586422 / 255.0, 115.59361427 / 255.0, 104.54012653 / 255.0]
	STD = [70.68188272 / 255.0, 68.27635443 / 255.0, 72.54505529 / 255.0]