Skip to content

Commit

Permalink
[lab01] Update code to tf 2.0, separate demo maskrcnn.
Browse files Browse the repository at this point in the history
  • Loading branch information
Arthur Douillard committed Nov 17, 2019
1 parent 66a4aae commit 134f6c1
Show file tree
Hide file tree
Showing 9 changed files with 265 additions and 252 deletions.
198 changes: 198 additions & 0 deletions labs/01_keras/Demo Mask RCNN.ipynb
View file
@@ -0,0 +1,198 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Instance Detection and Segmentation with Mask-RCNN\n",
"\n",
"[Mask RCNN](https://arxiv.org/abs/1703.06870) is a refinement of the [Faster RCNN](https://arxiv.org/abs/1506.01497) **object detection** model to also add support for **instance segmentation**.\n",
"\n",
"The following shows how to use a [Keras based implementation](https://github.com/matterport/Mask_RCNN) provided by matterport.com along with model parameters pretrained on the [COCO object detection dataset](http://cocodataset.org/).\n",
"\n",
"**WARNING**: The following requires to execute the companion `data_download.ipynb` notebook first.\n",
"\n",
"**WARNING**: For this notebook (and only this one), you'll need a tensorflow version under 2.0.0; create a new virtualenv and install requirements_tensorflow1.txt."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import tensorflow\n",
"import keras\n",
"\n",
"tensorflow.__version__, keras.__version__"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from maskrcnn import config\n",
"from maskrcnn import model as modellib\n",
"\n",
"\n",
"class InferenceCocoConfig(config.Config):\n",
" # Give the configuration a recognizable name\n",
" NAME = \"inference_coco\"\n",
"\n",
" # Number of classes (including background)\n",
" NUM_CLASSES = 1 + 80 # COCO has 80 classes\n",
"\n",
" # Set batch size to 1 since we'll be running inference on\n",
" # one image at a time. Batch size = GPU_COUNT * IMAGES_PER_GPU\n",
" GPU_COUNT = 1\n",
" IMAGES_PER_GPU = 1\n",
"\n",
"\n",
"config = InferenceCocoConfig()\n",
"model = modellib.MaskRCNN(mode=\"inference\", model_dir='maskrcnn/logs', config=config)\n",
"\n",
"# Load weights trained on MS-COCO\n",
"coco_model_file = \"mask_rcnn_coco.h5\"\n",
"model.load_weights(coco_model_file, by_name=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Class Names\n",
"\n",
"Index of the class in the list is its ID. For example, to get ID of the teddy bear class, use: `class_names.index('teddy bear')`\n",
"\n",
"`BG` stands for background."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# COCO Class names\n",
"class_names = ['BG', 'person', 'bicycle', 'car', 'motorcycle', 'airplane',\n",
" 'bus', 'train', 'truck', 'boat', 'traffic light',\n",
" 'fire hydrant', 'stop sign', 'parking meter', 'bench', 'bird',\n",
" 'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear',\n",
" 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie',\n",
" 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball',\n",
" 'kite', 'baseball bat', 'baseball glove', 'skateboard',\n",
" 'surfboard', 'tennis racket', 'bottle', 'wine glass', 'cup',\n",
" 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple',\n",
" 'sandwich', 'orange', 'broccoli', 'carrot', 'hot dog', 'pizza',\n",
" 'donut', 'cake', 'chair', 'couch', 'potted plant', 'bed',\n",
" 'dining table', 'toilet', 'tv', 'laptop', 'mouse', 'remote',\n",
" 'keyboard', 'cell phone', 'microwave', 'oven', 'toaster',\n",
" 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors',\n",
" 'teddy bear', 'hair drier', 'toothbrush']"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Run Object Detection"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from skimage.io import imread\n",
"\n",
"image = imread('webcam_shot.jpeg')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"image.shape"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from maskrcnn import visualize\n",
"import time\n",
"\n",
"# Run detection\n",
"tic = time.time()\n",
"results = model.detect([image], verbose=1)\n",
"toc = time.time()\n",
"print(\"Analyzed image in {:.3f}s\".format(toc - tic))\n",
"\n",
"# Visualize results\n",
"r = results[0]\n",
"for class_id, score in zip(r['class_ids'], r['scores']):\n",
" print(\"{}:\\t{:0.3f}\".format(class_names[class_id], score))\n",
"visualize.display_instances(image, r['rois'], r['masks'], r['class_ids'], \n",
" class_names, r['scores']);"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import os\n",
"import random\n",
"\n",
"# Load a random image from the images folder\n",
"image_folder = 'maskrcnn/images'\n",
"file_names = next(os.walk(image_folder))[2]\n",
"image = imread(os.path.join(image_folder, random.choice(file_names)))\n",
"\n",
"# Run detection\n",
"results = model.detect([image], verbose=1)\n",
"\n",
"# Visualize results\n",
"r = results[0]\n",
"visualize.display_instances(image, r['rois'], r['masks'], r['class_ids'], \n",
" class_names, r['scores'])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.5"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
39 changes: 22 additions & 17 deletions labs/01_keras/Intro Keras.ipynb
View file
Expand Up @@ -201,7 +201,7 @@
"outputs": [],
"source": [
"from tensorflow.keras.models import Sequential\n",
"from tensorflow.keras.layers import Dense\n",
"from tensorflow.keras.layers import Dense, Activation\n",
"from tensorflow.keras import optimizers\n",
"\n",
"input_dim = X_train.shape[1]\n",
Expand Down Expand Up @@ -298,7 +298,7 @@
"metadata": {},
"outputs": [],
"source": [
"rm -rf tensorboard_logs"
"!rm -rf tensorboard_logs"
]
},
{
Expand Down Expand Up @@ -367,6 +367,15 @@
"```"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"optimizers.SGD?"
]
},
{
"cell_type": "code",
"execution_count": null,
Expand Down Expand Up @@ -394,10 +403,7 @@
"\n",
"- Add another hidden layer and use the \"Rectified Linear Unit\" for each\n",
" hidden layer. Can you still train the model with Adam with its default global\n",
" learning rate?\n",
"\n",
"- Bonus: try the Adadelta optimizer (no learning rate to set).\n",
"\n"
" learning rate?"
]
},
{
Expand All @@ -413,7 +419,7 @@
"metadata": {},
"outputs": [],
"source": [
"# %load solutions/keras_adam_and_adadelta.py"
"# %load solutions/keras_adam.py"
]
},
{
Expand Down Expand Up @@ -467,7 +473,7 @@
"metadata": {},
"outputs": [],
"source": [
"type(predicted_labels_numpy)"
"type(predicted_labels_numpy), predicted_labels_numpy.shape"
]
},
{
Expand All @@ -493,7 +499,7 @@
"metadata": {},
"outputs": [],
"source": [
"type(predictions_tf)"
"type(predictions_tf), predictions_tf.shape"
]
},
{
Expand Down Expand Up @@ -703,8 +709,7 @@
"history = model.fit(X_train, Y_train, epochs=15, batch_size=32)\n",
"\n",
"plt.figure(figsize=(12, 4))\n",
"plt.plot(history.history['loss'], linestyle=linestyle,\n",
" label=\"Truncated Normal init\")\n",
"plt.plot(history.history['loss'], label=\"Truncated Normal init\")\n",
"plt.legend();"
]
},
Expand Down Expand Up @@ -750,25 +755,25 @@
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# %load solutions/keras_initializations.py"
]
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# %load solutions/keras_initializations_analysis.py"
"# %load solutions/keras_initializations.py"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
"source": [
"# %load solutions/keras_initializations_analysis.py"
]
}
],
"metadata": {
Expand All @@ -788,7 +793,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.4"
"version": "3.6.5"
},
"mimetype": "text/x-python",
"name": "python",
Expand Down

0 comments on commit 134f6c1

Please sign in to comment.