nn pic model preprocess note

最新推荐文章于 2024-08-19 08:22:45 发布

原创

最新推荐文章于 2024-08-19 08:22:45 发布 · 1.2w 阅读

0 ·

CC 4.0 BY-SA版权

本文探讨了深度学习中Inception和VGG模型的预处理步骤，并提到了一个关键的top2概念，同时包含了尝试应用的探索。

inception 预处理

https://github.com/fchollet/keras/blob/master/keras/applications/inception_v3.py

def preprocess_input(x):
    x /= 255.
    x -= 0.5
    x *= 2.
    return x

vgg 预处理

    if len(x.shape) == 3:
        #print("[INFO] channel:{}".format(data_format))
        if data_format == 'channels_first':
            x = x.transpose(2, 0, 1)
    elif len(x.shape) == 2:
        if data_format == 'channels_first':
            x = x.reshape((1, x.shape[0], x.shape[1]))
        else:
            x = x.reshape((x.shape[0], x.shape[1], 1))
    else:
        raise ValueError('Unsupported image shape: ', x.shape)

im[:,:,0] -= 103.939
im[:,:,1] -= 116.779
im[:,:,2] -= 123.68

top2

这里写图片描述

try

_________________________________________________________________
Layer (type)                 Output Shape              Param #
================================================