├── Action_detection.ipynb
├── README.md
├── images
    ├── 1.jpg
    ├── 2.jpg
    ├── 3.jpg
    ├── 4.jpg
    ├── 5.jpg
    └── action_net.jpg
└── model_class.json


/README.md:
--------------------------------------------------------------------------------
  1 | # Action-Detection(Action-Net)
  2 | Action-Net is a dataset containing images of 16 different human actions.
  3 | <br><br>
  4 | <img src="images/action_net.jpg" />
  5 | <hr>
  6 | <b>Action-Net</b> is a dataset containing images of human actions , collected in order to ensure that machine learning systems can be trained
  7 |  to understand human actions, gestures and activities. This is part of <a href="https://deepquestai.com" >DeepQuest AI</a>'s to train machine learning systems to 
  8 |   perceive, understand and act accordingly in solving problems in any environment they are deployed. <br><br>
  9 | 
 10 |   This is the first release of the Action-Net dataset. It contains 19,200 images that span cover 16 classes. The classes
 11 |   included in this release are: <br><br>
 12 | 
 13 |   - <b> Calling </b> <br>
 14 |   - <b> Clapping </b> <br>
 15 |   - <b> Cycling </b> <br>
 16 |   - <b> Dancing </b> <br>
 17 |   - <b> Drinking </b> <br>
 18 |   - <b> Eating </b> <br>
 19 |   - <b> Fighting </b> <br>
 20 |   - <b> Hugging </b> <br>
 21 |   - <b> Kissing </b> <br>
 22 |   - <b> Laughing </b> <br>
 23 |   - <b> Listening to Music </b> <br>
 24 |   - <b> Running </b> <br>
 25 |   - <b> Sitting </b> <br>
 26 |   - <b> Sleeping </b> <br>
 27 |   - <b> Texting </b> <br>
 28 |   - <b> Using Laptop </b> <br>
 29 |   
 30 | 
 31 |   There are <b>1,200 images</b> for each category, with <b>1000 images for trainings </b> and <b>200 images for testing</b> . We are working on adding more
 32 |    categories in the future and will continue to improve the dataset.
 33 |   <br><br> <br>
 34 | 
 35 |   <b>>>> DOWNLOAD, TRAINING AND PREDICTION: </b> <br><br>
 36 |  The <b>Action-Net</b> dataset is provided for download in the <b>release</b> section of this repository.
 37 |  You can download the dataset via the link below.<br><br> <a href="https://github.com/OlafenwaMoses/Action-Net/releases/download/v1/action_net_v1.zip" >https://github.com/OlafenwaMoses/Action-Net/releases/download/v1/action_net_v1.zip</a>  <br><br>
 38 | 
 39 |  We have also provided a python codebase to download the images, train <b>ResNet50</b> on the images
 40 |   and perform prediction using a pretrained model (also using <b>ResNet50</b>) provided in the release section of this repository.
 41 |   The python codebase is contained in the <b><a href="action_net.py" >action_net.py</a></b> file and the model class labels for prediction is also provided the 
 42 |   <b><a href="model_class.json" >model_class.json</a></b>. The pretrained <b>ResNet50</b> model is available for download via the link below. <br><br> 
 43 |   <b><a href="https://github.com/OlafenwaMoses/Action-Net/releases/download/v1/action_net_ex-060_acc-0.745313.h5" >https://github.com/OlafenwaMoses/Action-Net/releases/download/v1/action_net_ex-060_acc-0.745313.h5</a></b><br>
 44 |   <br>
 45 |    This pre-trained model was trained for **60 epochs** only, but it achieved over **74%** accuracy on 3200 test images. You can see the prediction results on new images that were not part of the dataset in the **Prediction Results** section below. More experiments will enhance the accuracy of the model.
 46 | <br>
 47 | Running the experiment or prediction requires that you have **Tensorflow**, and **Keras**, **OpenCV** and **ImageAI** installed. You can install this dependencies via the commands below.
 48 | 
 49 | <br><span><b>- Tensorflow 1.4.0 (and later versions)  </b>      <a href="https://www.tensorflow.org/install/install_windows" style="text-decoration: none;" > Install</a></span> or install via pip <pre> pip3 install --upgrade tensorflow </pre> 
 50 |        
 51 |   <span><b>- OpenCV  </b>        <a href="https://pypi.python.org/pypi/opencv-python" style="text-decoration: none;" >Install</a></span> or install via pip <pre> pip3 install opencv-python </pre> 
 52 |        
 53 |    <span><b>- Keras 2.x  </b>     <a href="https://keras.io/#installation" style="text-decoration: none;" >Install</a></span> or install via pip <pre> pip3 install keras </pre> 
 54 |   
 55 |    <span><b>- ImageAI 2.0.3  </b>  
 56 |    <span>      <pre>pip3 install imageai </pre></span> <br><br> <br>
 57 | 
 58 | 
 59 | 
 60 | 
 61 |   <img src="images/1.jpg" />
 62 | <pre>
 63 | eating  :  100.0
 64 | drinking  :  3.92037860508232e-09
 65 | using-laptop  :  6.944534465709584e-11
 66 | calling  :  5.7910951424891555e-12
 67 | </pre>
 68 | 
 69 | <hr>
 70 | <br>
 71 | <img src="images/2.jpg" />
 72 | <pre>
 73 | eating  :  99.44907426834106
 74 | drinking  :  0.5508399568498135
 75 | using-phone  :  5.766927415606915e-05
 76 | sitting  :  1.1222620344142342e-05
 77 | </pre>
 78 | 
 79 | <hr>
 80 | <br>
 81 | 
 82 | <img src="images/3.jpg" />
 83 | <pre>
 84 | fighting  :  99.97442364692688
 85 | running  :  0.01658390392549336
 86 | dancing  :  0.008970857743406668
 87 | sitting  :  7.210289965087213e-06
 88 | </pre>
 89 | 
 90 | <hr>
 91 | <br>
 92 | 
 93 | <img src="images/4.jpg" />
 94 | <pre>
 95 | laughing  :  99.99998807907104
 96 | clapping  :  1.3144966715117334e-05
 97 | calling  :  4.0294068526236515e-06
 98 | eating  :  4.981405066217803e-07
 99 | </pre>
100 | 
101 | 
102 | <hr>
103 | <br>
104 | 
105 | <img src="images/5.jpg" />
106 | <pre>
107 | running  :  99.99852180480957
108 | calling  :  0.0009251662959286477
109 | listening-to-music  :  0.0002909338491008384
110 | cycling  :  0.00024121977730828803
111 | </pre>
112 | 
113 | 
114 | <br>
115 | 


--------------------------------------------------------------------------------
/images/1.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/aasu14/Action-Detection/25282f8b284f97189dd0d3bb48e2b022099b6738/images/1.jpg


--------------------------------------------------------------------------------
/images/2.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/aasu14/Action-Detection/25282f8b284f97189dd0d3bb48e2b022099b6738/images/2.jpg


--------------------------------------------------------------------------------
/images/3.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/aasu14/Action-Detection/25282f8b284f97189dd0d3bb48e2b022099b6738/images/3.jpg


--------------------------------------------------------------------------------
/images/4.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/aasu14/Action-Detection/25282f8b284f97189dd0d3bb48e2b022099b6738/images/4.jpg


--------------------------------------------------------------------------------
/images/5.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/aasu14/Action-Detection/25282f8b284f97189dd0d3bb48e2b022099b6738/images/5.jpg


--------------------------------------------------------------------------------
/images/action_net.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/aasu14/Action-Detection/25282f8b284f97189dd0d3bb48e2b022099b6738/images/action_net.jpg


--------------------------------------------------------------------------------
/model_class.json:
--------------------------------------------------------------------------------
 1 | {
 2 |     "0" : "calling",
 3 |     "1" : "clapping",
 4 |     "2" : "cycling",
 5 |     "3" : "dancing",
 6 |     "4" : "drinking",
 7 |     "5" : "eating",
 8 |     "6" : "fighting",
 9 |     "7" : "hugging",
10 |     "8" : "kissing",
11 |     "9" : "laughing",
12 |     "10" : "listening-to-music",
13 |     "11" : "running",
14 |     "12" : "sitting",
15 |     "13" : "sleeping",
16 |     "14" : "using-laptop",
17 |     "15" : "using-phone"
18 | }


--------------------------------------------------------------------------------