垃圾图像分类 ResNet34 python

数据下载链接 https://pan.baidu.com/s/1wr3h2Wc720uqUeIroTCIJA百度网盘为您提供文件的网络备份、同步和分享服务。空间大、速度快、安全稳固，支持教育网加速，支持手机端。注册使用百度网盘即可享受免费存储空间https://pan.baidu.com/s/1wr3h2Wc720uqUeIroTCIJA

提取码:mqic

污染是回收行业的一个巨大问题，可以通过自动化废物分类来缓解。只是为了踢球，我想我会尝试制作一个图像分类器的原型来对垃圾和可回收物进行分类 – 这个分类器可以在光学分拣系统中得到应用。

waste_types = ['hazardous_waste_dry_battery', 'hazardous_waste_expired_drugs', 'hazardous_waste_ointment', 'kitchen_waste_bone', 'kitchen_waste_eggshell', 'kitchen_waste_fish_bone', 'kitchen_waste_fruit_peel', 'kitchen_waste_meal', 'kitchen_waste_pulp', 'kitchen_waste_tea', 'kitchen_waste_vegetable', 'other_garbage_bamboo_chopsticks', 'other_garbage_cigarette', 'other_garbage_fast_food_box', 'other_garbage_flowerpot', 'other_garbage_soiled_plastic', 'other_garbage_toothpick', 'recyclables_anvil', 'recyclables_bag', 'recyclables_bottle', 'recyclables_can', 'recyclables_cardboard', 'recyclables_cosmetic_bottles', 'recyclables_drink_bottle', 'recyclables_edible_oil_barrel', 'recyclables_glass_cup', 'recyclables_metal_food_cans', 'recyclables_old_clothes', 'recyclables_paper_bags', 'recyclables_pillow', 'recyclables_plastic_bowl', 'recyclables_plastic_hanger', 'recyclables_plug_wire', 'recyclables_plush_toys', 'recyclables_pot', 'recyclables_powerbank', 'recyclables_seasoning_bottle', 'recyclables_shampoo_bottle', 'recyclables_shoes', 'recyclables_toys']

一些基本的准备

%reload_ext autoreload %autoreload 2 %matplotlib inline %config InlineBackend.figure_format = 'retina'

from fastai.vision import * from fastai.metrics import error_rate from pathlib import Path from glob2 import glob from sklearn.metrics import confusion_matrix import pandas as pd import numpy as np import os import zipfile as zf import shutil import re import seaborn as sns

files = zf.ZipFile("dataset-resized.zip",'r') files.extractall() files.close()

解压缩后，数据集调整大小的文件夹有40个子文件夹：

垃圾图像分类 ResNet34 python

os.listdir(os.path.join(os.getcwd(),"dataset-resized"))

 helper functions splits indices for a folder into train, validation, and test indices with random sampling input: folder path output: train, valid, and test indices def split_indices(folder,seed1,seed2): n = len(os.listdir(folder)) full_set = list(range(1,n+1)) train indices random.seed(seed1) train = random.sample(list(range(1,n+1)),int(.5*n)) temp remain = list(set(full_set)-set(train)) separate remaining into validation and test random.seed(seed2) valid = random.sample(remain,int(.5*len(remain))) test = list(set(remain)-set(valid)) return(train,valid,test) gets file names for a particular type of trash, given indices input: waste category and indices output: file names def get_names(waste_type,indices): file_names = [waste_type+str(i)+".jpg" for i in indices] return(file_names) moves group of source files to another folder input: list of source files and destination folder no output def move_files(source_files,destination_folder): for file in source_files: shutil.move(file,destination_folder)

之后训练集和验证集里面各有四十个文件夹

垃圾图像分类 ResNet34 python

 paths will be train/cardboard, train/glass, etc... subsets = ['train','valid'] waste_types = waste_types = ['hazardous_waste_dry_battery', 'hazardous_waste_expired_drugs', 'hazardous_waste_ointment', 'kitchen_waste_bone', 'kitchen_waste_eggshell', 'kitchen_waste_fish_bone', 'kitchen_waste_fruit_peel', 'kitchen_waste_meal', 'kitchen_waste_pulp', 'kitchen_waste_tea', 'kitchen_waste_vegetable', 'other_garbage_bamboo_chopsticks', 'other_garbage_cigarette', 'other_garbage_fast_food_box', 'other_garbage_flowerpot', 'other_garbage_soiled_plastic', 'other_garbage_toothpick', 'recyclables_anvil', 'recyclables_bag', 'recyclables_bottle', 'recyclables_can', 'recyclables_cardboard', 'recyclables_cosmetic_bottles', 'recyclables_drink_bottle', 'recyclables_edible_oil_barrel', 'recyclables_glass_cup', 'recyclables_metal_food_cans', 'recyclables_old_clothes', 'recyclables_paper_bags', 'recyclables_pillow', 'recyclables_plastic_bowl', 'recyclables_plastic_hanger', 'recyclables_plug_wire', 'recyclables_plush_toys', 'recyclables_pot', 'recyclables_powerbank', 'recyclables_seasoning_bottle', 'recyclables_shampoo_bottle', 'recyclables_shoes', 'recyclables_toys'] create destination folders for data subset and waste type for subset in subsets: for waste_type in waste_types: folder = os.path.join('data',subset,waste_type) if not os.path.exists(folder): os.makedirs(folder) if not os.path.exists(os.path.join('data','test')): os.makedirs(os.path.join('data','test')) move files to destination folders for each waste type for waste_type in waste_types: source_folder = os.path.join('train',waste_type) train_ind, valid_ind, test_ind = split_indices(source_folder,1,1) move source files to train train_names = get_names(waste_type,train_ind) train_source_files = [os.path.join(source_folder,name) for name in train_names] train_dest = "data/train/"+waste_type move_files(train_source_files,train_dest) move source files to valid valid_names = get_names(waste_type,valid_ind) valid_source_files = [os.path.join(source_folder,name) for name in valid_names] valid_dest = "data/valid/"+waste_type move_files(valid_source_files,valid_dest) move source files to test test_names = get_names(waste_type,test_ind) test_source_files = [os.path.join(source_folder,name) for name in test_names] I use data/test here because the images can be mixed up move_files(test_source_files,"data/test")

为了可重复性，我将两个随机样本的种子设置为 1。现在数据已经组织好，我们可以开始模型训练了。

 get a path to the folder with images path = Path(os.getcwd())/"data" path

out:PosixPath('/home/jupyter/data')

tfms = get_transforms(do_flip=True,flip_vert=True) data = ImageDataBunch.from_folder(path,test="test",ds_tfms=tfms,bs=16) data #可以把data打印出来看看

data.show_batch(rows=4,figsize=(10,8)) #显示图片

训练模型！

什么是resnet34？
残差神经网络是具有很多层的卷积神经网络 (CNN)。特别是，resnet34 是一个 34 层的 CNN，已经在 ImageNet 数据库上进行了预训练。预训练的 CNN 将在新的图像分类任务上表现得更好，因为它已经学习了一些视觉特征并且可以将这些知识转移（因此是转移学习）。

由于它们能够描述更多的复杂性，理论上深度神经网络在训练数据上应该比浅层网络表现得更好。但实际上，深度神经网络在经验上的表现往往比浅层神经网络差。

创建 Resnets 是为了使用一种称为快捷连接的黑客来规避这个故障。如果某个层中的某些节点具有次优值，则可以调整权重和偏差；如果一个节点是最优的（它的残差为 0），为什么不把它放在一边？仅根据需要对节点进行调整（当存在非零残差时）。

当需要调整时，快捷连接应用恒等函数将信息传递给后续层。这在可能的情况下缩短了神经网络，并允许 resnet 具有深层架构并表现得更像浅层神经网络。 resnet34中的34只是指层数。

Anand Saha 在这里给出了更深入的解释。

learn = create_cnn(data,models.resnet34,metrics=error_rate) learn.model

learn.lr_find(start_lr=1e-6,end_lr=1e1) learn.recorder.plot()

垃圾图像分类 ResNet34 python

learn.fit_one_cycle(20,max_lr=5.13e-03)

垃圾图像分类 ResNet34 python

我的模型运行了 20 个 epoch。这种拟合方法的酷炫之处在于，学习率随着每个 epoch 的增长而降低，让我们越来越接近最佳状态。在 8.6% 时，验证错误看起来非常好……让我们看看它在测试数据上的表现如何。

首先，我们可以看看哪些图像分类错误最多。

interp = ClassificationInterpretation.from_learner(learn) losses,idxs = interp.top_losses() interp.plot_top_losses(9, figsize=(15,11))

doc(interp.plot_top_losses) interp.plot_confusion_matrix(figsize=(12,12), dpi=60)

interp.most_confused(min_val=2)

现在就到了最激动人心的时刻

注意：learner.predict() 只对单个图像进行预测，而 learner.get_preds() 对一组图像进行预测。我强烈建议阅读文档以了解有关 predict() 和 get_preds() 的更多信息。

preds = learn.get_preds(ds_type=DatasetType.Test)

get_preds(ds_type) 中的 ds_type 参数采用 DataSet 参数。示例值为 DataSet.Train、DataSet.Valid 和 DataSet.Test。我提到这一点是因为我错误地传入了实际数据 (learn.data.test_ds)，这给了我错误的输出并且花了很长时间进行调试。不要犯这个错误！不要传入数据——传入数据集类型！

print(preds[0].shape) preds[0]

结果就在yhat里面！

 saves the index (0 to 5) of most likely (max) predicted class for each image max_idxs = np.asarray(np.argmax(preds[0],axis=1)) yhat = [] for max_idx in max_idxs: yhat.append(data.classes[max_idx]) yhat

…..

recyclables_plush_toys kitchen_waste_eggshell recyclables_edible_oil_barrel hazardous_waste_expired_drugs kitchen_waste_eggshell recyclables_shoes recyclables_plug_wire kitchen_waste_vegetable recyclables_shoes recyclables_toys recyclables_seasoning_bottle recyclables_bag kitchen_waste_fruit_peel other_garbage_cigarette recyclables_can recyclables_anvil other_garbage_cigarette recyclables_shoes recyclables_paper_bags kitchen_waste_fish_bone other_garbage_bamboo_chopsticks other_garbage_bamboo_chopsticks other_garbage_flowerpot recyclables_bottle kitchen_waste_vegetable kitchen_waste_pulp recyclables_edible_oil_barrel recyclables_plastic_bowl other_garbage_fast_food_box recyclables_pot recyclables_cardboard recyclables_glass_cup recyclables_plastic_hanger recyclables_paper_bags recyclables_seasoning_bottle kitchen_waste_bone recyclables_seasoning_bottle recyclables_powerbank recyclables_drink_bottle kitchen_waste_fruit_peel recyclables_seasoning_bottle recyclables_powerbank recyclables_plush_toys recyclables_plush_toys recyclables_seasoning_bottle recyclables_bottle recyclables_edible_oil_barrel recyclables_edible_oil_barrel recyclables_plastic_bowl recyclables_metal_food_cans other_garbage_cigarette hazardous_waste_expired_drugs kitchen_waste_fish_bone recyclables_shoes kitchen_waste_pulp recyclables_plastic_bowl recyclables_powerbank

….

或者可以看另外两篇图像分类的

最简单的直接用训练好的模型

https://blog.csdn.net/long_songs/article/details/122095136https://blog.csdn.net/long_songs/article/details/122095136