Open3D面向机器学习的扩展库 / 开普饭

Open3D-ML是Open3D的一个扩展，用于3D机器学习任务。它建立在Open3D核心库之上，并通过机器学习工具对其进行扩展，以进行3D数据处理。此repo集中于语义点云分割等应用程序，并提供可应用于常见任务的预训练模型以及用于训练的流程。

Open3D-ML与TensorFlow和PyTorch一起工作，可以轻松地集成到现有项目中，还可以提供独立于ML框架的通用功能，如数据可视化。

安装

Open3D-ML集成在Open3D v0.11+python发行版中，并与以下版本的ML框架兼容

* PyTorch 1.6

* TensorFlow 2.3

* CUDA 10.1 (On GNU/Linux x86_64, optional)

安装Open3D

# make sure you have the latest pip versionpip install --upgrade pip# install open3dpip install open3d

要安装Pythorch或TensorFlow的兼容版本，需要使用相应的需求文件：

# To install a compatible version of TensorFlowpip install -r requirements-tensorflow.txt# To install a compatible version of PyTorch with CUDApip install -r requirements-torch-cuda.txt

测试安装

# with PyTorch$ python -c "import open3d.ml.torch as ml3d"# or with TensorFlow$ python -c "import open3d.ml.tf as ml3d"

如果需要使用不同版本的ML框架或CUDA，可以从源代码重新构建Open3D。

使用教程

读取数据集

dataset命名空间包含用于读取公共数据集的类。这里我们读取SemanticKITTI数据集并将其可视化。

import open3d.ml.torch as ml3d # or open3d.ml.tf as ml3d

# construct a dataset by specifying dataset_pathdataset = ml3d.datasets.SemanticKITTI(dataset_path='/path/to/SemanticKITTI/')

# get the 'all' split that combines training, validation and test setall_split = dataset.get_split('all')

# print the attributes of the first datumprint(all_split.get_attr(0))

# print the shape of the first point cloudprint(all_split.get_data(0)['point'].shape)

# show the first 100 frames using the visualizervis = ml3d.vis.Visualizer()

vis.visualize_dataset(dataset, 'all', indices=range(100))import open3d.ml.torch as ml3d # or open3d.ml.tf as ml3d

# construct a dataset by specifying dataset_pathdataset = ml3d.datasets.SemanticKITTI(dataset_path='/path/to/SemanticKITTI/')

# get the 'all' split that combines training, validation and test setall_split = dataset.get_split('all')# print the attributes of the first datumprint(all_split.get_attr(0))# print the shape of the first point cloudprint(all_split.get_data(0)['point'].shape)# show the first 100 frames using the visualizervis = ml3d.vis.Visualizer()vis.visualize_dataset(dataset, 'all', indices=range(100))

加载配置文件

模型、数据集和流程的配置存储在ml3d/Configs中。用户还可以构建自己的yaml文件来记录他们的定制配置。下面是一个读取配置文件并从中构造模块的示例。

import open3d.ml as _ml3dimport open3d.ml.torch as ml3d # or open3d.ml.tf as ml3d

framework = "torch" # or tfcfg_file = "ml3d/configs/randlanet_semantickitti.yml"cfg = _ml3d.utils.Config.load_from_file(cfg_file)

# fetch the classes by the namePipeline = _ml3d.utils.get_module("pipeline", cfg.pipeline.name, framework)Model = _ml3d.utils.get_module("model", cfg.model.name, framework)Dataset = _ml3d.utils.get_module("dataset", cfg.dataset.name)

# use the arguments in the config file to construct the instances cfg.dataset['dataset_path'] = "/path/to/your/dataset"dataset = Dataset(cfg.dataset.pop('dataset_path', None), **cfg.dataset)model = Model(**cfg.model)pipeline = Pipeline(model, dataset, **cfg.pipeline)

运行一个预先训练过的模型

在上一个例子的基础上，我们可以用一个预先训练的语义分割模型实例化一个算法，并在数据集的点云上运行它。查看模型集合以获取预训练模型的权重。

import osimport open3d.ml as _ml3dimport open3d.ml.torch as ml3d

cfg_file = "ml3d/configs/randlanet_semantickitti.yml"cfg = _ml3d.utils.Config.load_from_file(cfg_file)

model = ml3d.models.RandLANet(**cfg.model)cfg.dataset['dataset_path'] = "/path/to/your/dataset"dataset = ml3d.datasets.SemanticKITTI(cfg.dataset.pop('dataset_path', None), **cfg.dataset)pipeline = ml3d.pipelines.SemanticSegmentation(model, dataset=dataset, device="gpu", **cfg.pipeline)

# download the weights.ckpt_folder = "./logs/"os.makedirs(ckpt_folder, exist_ok=True)ckpt_path = ckpt_folder + "randlanet_semantickitti_202009090354utc.pth"randlanet_url = "https://storage.googleapis.com/open3d-releases/model-zoo/randlanet_semantickitti_202009090354utc.pth"if not os.path.exists(ckpt_path): cmd = "wget {} -O {}".format(randlanet_url, ckpt_path) os.system(cmd) # load the parameters.pipeline.load_ckpt(ckpt_path=ckpt_path)

test_split = dataset.get_split("test")data = test_split.get_data(0)

# run inference on a single example.# returns dict with 'predict_labels' and 'predict_scores'.result = pipeline.run_inference(data)

# evaluate performance on the test set; this will write logs to './logs'.pipeline.run_test()

用户还可以使用预定义的脚本来加载预先训练的权重并运行测试。

训练模型

与推理类似，流程中提供了一个在数据集上训练模型的接口。

# use a cache for storing the results of the preprocessing (default path is './logs/cache')dataset = ml3d.datasets.SemanticKITTI(dataset_path='/path/to/SemanticKITTI/', use_cache=True)

# create the model with random initialization.model = RandLANet()

pipeline = SemanticSegmentation(model=model, dataset=dataset, max_epoch=100)

# prints training progress in the console.pipeline.run_train()

有关更多示例，请参见examples/和scripts/目录。

使用预定义脚本

scripts/semseg.py 提供了一个简单的数据集评估接口。准确地定义模型，避免了定义具体模型的麻烦。

python scripts/semseg.py {tf/torch} -c <path-to-config> --<extra args>

注意， extra args 将优先于配置文件中的相同参数。因此，在启动脚本时，可以将其作为命令行参数传递，而不是更改配置文件中的param。

例如：

# Launch training for RandLANet on SemanticKITTI with torch. python scripts/semseg.py torch -c ml3d/configs/randlanet_semantickitti.yml --dataset.dataset_path <path-to-dataset> --dataset.use_cache True

# Launch testing for KPConv on Toronto3D with tensorflow. python scripts/semseg.py tf -c ml3d/configs/kpconv_toronto3d.yml --split test --dataset.dataset_path <path-to-dataset> --model.ckpt_path <path-to-checkpoint>

要获得进一步的帮助，可运行python脚本 python scripts/semseg.py --help

ML库结构

Open3D-ML的核心部分位于ml3d子文件夹中，该子文件夹被集成到ML命名空间中的Open3D中。除了核心部分之外，目录示例和脚本还提供了支持脚本，用于开始在数据集上设置训练流程或运行网络。

├─ docs # Markdown and rst files for documentation

├─ examples # Place for example scripts and notebooks

├─ ml3d # Package root dir that is integrated in open3d

├─ configs # Model configuration files

├─ datasets # Generic dataset code; will be integratede as open3d.ml.{tf,torch}.datasets

├─ utils # Framework independent utilities; available as open3d.ml.{tf,torch}.utils

├─ vis # ML specific visualization functions

├─ tf # Directory for TensorFlow specific code. same structure as ml3d/torch.

│ # This will be available as open3d.ml.tf

├─ torch # Directory for PyTorch specific code; available as open3d.ml.torch

├─ dataloaders # Framework specific dataset code, e.g. wrappers that can make use of the

│ # generic dataset code.

├─ models # Code for models

├─ modules # Smaller modules, e.g., metrics and losses

├─ pipelines # Pipelines for tasks like semantic segmentation

├─ scripts # Demo scripts for training and dataset download scripts

任务和算法

分割

对于语义分割的任务，我们使用mIoU（mean intersection-over-union）来衡量不同方法在所有类上的性能。下表显示了分段任务的可用模型和数据集以及各自的分数。每个分数链接到各自的权重文件。