buseyaren/installation-guide-of-maskrcnn

Mask R-CNN creates a high-quality segmentation mask in addition to the Faster R-CNN network. In addition to class labels and scores, a segmentation mask is created for the objects detected by this neural network. In this repository, using Anaconda prompt step by step Mask R-CNN setup is shown.

Step By Step Mask RCNN Installation

Attention❗️

• Compatible Python Version: python==3.6.12

• IDE: Anaconda Cloud & Conda Prompt

  -Anaconda Cloud: https://www.anaconda.com

🔺 Step 1: Compatible with Python 3.6 version, a virtual environment named maskrcnn is created in conda prompt.

conda create -n maskrcnn python=3.6.12

🔺 Step 2: The maskrcnn virtual environment is activated.

conda activate maskrcnn

🔺 Step 3: The Mask RCNN published by Matterport is cloned from the GitHub repository.

git clone https://github.com/matterport/Mask_RCNN.git

🔺 Step 4: Mask RCNN must be installed in the requirements.txt file located in the GitHub store. The requirements.txt file will load the libraries needed for your project in batch.

pip install -r requirements.txt

Dependencies

numpy, scipy, cython, h5py, Pillow, scikit-image, tensorflow==1.14.0 keras==2.0.8, jupyter or (tensorflow==1.15.0 keras==2.2.5)

For GPU: tensorflow-gpu:1.15.0, keras:2.2.5 For CPU: tensorflow:1.14.0, keras:2.0.8, h5py:2.10.0

🔺 Step 5: Download the pre-trained weights from https://github.com/matterport/Mask_RCNN/releases.

Download the file mask_rcnn_balloon.h5 from Mask_RCNN_2.1 file and mask_rcnn_coco.h5 model from Mask_RCNN_2.0 file. These 2 models should be placed in the samples folder.

Attention❗️

If the TensorFlow and Keras versions have landed in high versions, you can make a specific installation with the following commands.

🔺 Step 6: Running the setup.py file.

python setup.py install

🔺 Step 7: Loading the pycocotols module.

pip install git+https://github.com/philferriere/cocoapi.git#subdirectory=PythonAPI

🔺 Step 8: Let's run it on the Jupyter notebook.

jupyter notebook

A view from the project:

installation-guide-of-maskrcnn's Issues

Im not able to run the demo after following your steps

I have done exactly what you described in the readme. I have installed 1.15.0 tensorflow with gpu and then keras 2.2.5.
I have done everything else. However when i run the jupyter notebook in the virtual environment, I get the following error

ttributeError                            Traceback (most recent call last)
<ipython-input-7-ebe7095df7bb> in <module>
     14 sys.path.append(ROOT_DIR)  # To find local version of the library
     15 from mrcnn import utils
---> 16 import mrcnn.model as modellib
     17 from mrcnn import visualize
     18 # Import COCO config

~/anaconda3/envs/maskrcnn/lib/python3.6/importlib/_bootstrap.py in _find_and_load(name, import_)

~/anaconda3/envs/maskrcnn/lib/python3.6/importlib/_bootstrap.py in _find_and_load_unlocked(name, import_)

~/anaconda3/envs/maskrcnn/lib/python3.6/importlib/_bootstrap.py in _load_unlocked(spec)

~/anaconda3/envs/maskrcnn/lib/python3.6/importlib/_bootstrap.py in _load_backward_compatible(spec)

~/anaconda3/envs/maskrcnn/lib/python3.6/site-packages/mask_rcnn-2.1-py3.6.egg/mrcnn/model.py in <module>
     18 import numpy as np
     19 import tensorflow as tf
---> 20 import keras
     21 import keras.backend as K
     22 import keras.layers as KL

~/.local/lib/python3.6/site-packages/keras/__init__.py in <module>
      1 from __future__ import absolute_import
      2 
----> 3 from . import utils
      4 from . import activations
      5 from . import applications

~/.local/lib/python3.6/site-packages/keras/utils/__init__.py in <module>
     22 from .layer_utils import get_source_inputs
     23 from .layer_utils import print_summary
---> 24 from .vis_utils import model_to_dot
     25 from .vis_utils import plot_model
     26 from .np_utils import to_categorical

~/.local/lib/python3.6/site-packages/keras/utils/vis_utils.py in <module>
      5 
      6 import os
----> 7 from ..models import Model
      8 from ..layers.wrappers import Wrapper
      9 

~/.local/lib/python3.6/site-packages/keras/models.py in <module>
      8 from .utils.generic_utils import has_arg
      9 from .utils.generic_utils import to_list
---> 10 from .engine.input_layer import Input
     11 from .engine.input_layer import InputLayer
     12 from .engine.training import Model

~/.local/lib/python3.6/site-packages/keras/engine/__init__.py in <module>
      1 # note: `Node` is an internal class,
      2 # it isn't meant to be used by Keras users.
----> 3 from .input_layer import Input
      4 from .input_layer import InputLayer
      5 from .base_layer import InputSpec

~/.local/lib/python3.6/site-packages/keras/engine/input_layer.py in <module>
      5 from __future__ import division
      6 
----> 7 from .base_layer import Layer
      8 from .base_layer import Node
      9 from .. import backend as K

~/.local/lib/python3.6/site-packages/keras/engine/base_layer.py in <module>
      9 
     10 from .. import backend as K
---> 11 from .. import initializers
     12 from ..utils.layer_utils import count_params
     13 from ..utils.generic_utils import has_arg

~/.local/lib/python3.6/site-packages/keras/initializers/__init__.py in <module>
    122 # from ALL_OBJECTS. We make no guarantees as to whether these objects will
    123 # using their correct version.
--> 124 populate_deserializable_objects()
    125 globals().update(LOCAL.ALL_OBJECTS)
    126 

~/.local/lib/python3.6/site-packages/keras/initializers/__init__.py in populate_deserializable_objects()
     47 
     48   LOCAL.ALL_OBJECTS = {}
---> 49   LOCAL.GENERATED_WITH_V2 = tf.__internal__.tf2.enabled()
     50 
     51   # Compatibility aliases (need to exist in both V1 and V2).

~/anaconda3/envs/maskrcnn/lib/python3.6/site-packages/tensorflow_core/python/util/module_wrapper.py in __getattr__(self, name)
    191   def __getattr__(self, name):
    192     try:
--> 193       attr = getattr(self._tfmw_wrapped_module, name)
    194     except AttributeError:
    195       if not self._tfmw_public_apis:

AttributeError: module 'tensorflow._api.v1.compat.v2' has no attribute '__internal__'

is creating a virtual environment necessary or optional?