Log In Sign Up
Models:
Wanda Brown
WandaB/
Automatic-Identification
0
Downloads: 1
Diff of /README.md [d34869] .. [45a3e1]

Switch to unified view

a/README.md b/README.md
1 
<h2 align="center">Automatic Identification and Counting of Blood Cells<h2>
 1 
<h2 align="center">Automatic Identification and Counting of Blood Cells</h2>
2 

  2
  





  3
  
[![GitHub stars](https://img.shields.io/github/stars/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/stargazers)

  3
  
[![GitHub stars](https://img.shields.io/github/stars/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/stargazers)






  4
  
[![GitHub forks](https://img.shields.io/github/forks/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/network)

  4
  
[![GitHub forks](https://img.shields.io/github/forks/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/network)






  5
  
[![GitHub issues](https://img.shields.io/github/issues/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/issues)

  5
  
[![GitHub issues](https://img.shields.io/github/issues/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/issues)






  6
  
[![Version](https://img.shields.io/badge/version-2.1-green.svg?longCache=true&style=flat)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)

  6
  
[![Version](https://img.shields.io/badge/version-2.1-green.svg?longCache=true&style=flat)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)






  7
  
[![GitHub license](https://img.shields.io/github/license/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/blob/master/LICENSE)

  7
  
[![GitHub license](https://img.shields.io/github/license/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells)](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/blob/master/LICENSE)






  8
  
<!--<img src="https://user-images.githubusercontent.com/37298971/123714340-f8d70800-d82a-11eb-9742-042a5d9334a1.png" width="28">-->

  8
  
<!--<img src="https://user-images.githubusercontent.com/37298971/123714340-f8d70800-d82a-11eb-9742-042a5d9334a1.png" width="28">-->





  9
  

  9
  





  10
  
## Dataset

  10
  
## Dataset





  11
  

  11
  





  12
  
The [```Complete Blood Count (CBC) Dataset```](https://github.com/MahmudulAlam/Complete-Blood-Cell-Count-Dataset) has

  12
  
The [```Complete Blood Count (CBC) Dataset```](https://github.com/MahmudulAlam/Complete-Blood-Cell-Count-Dataset) has






  13
  
been used for automatic identification and counting of blood cells. Download the dataset, unzip and put

  13
  
been used for automatic identification and counting of blood cells. Download the dataset, unzip and put






  14
  
the ```Training```, ```Testing```, and ```Validation```folders in the working directory.

  14
  
the ```Training```, ```Testing```, and ```Validation```folders in the working directory.





  15
  

  15
  





  16
  
## Requirements

  16
  
## Requirements





  17
  

  17
  





  18
  
![requirements](https://img.shields.io/badge/Python-3.6-3480eb.svg?longCache=true&style=flat&logo=python)

  18
  
![requirements](https://img.shields.io/badge/Python-3.6-3480eb.svg?longCache=true&style=flat&logo=python)






  19
  
![requirements](https://img.shields.io/badge/Python-3.7-3480eb.svg?longCache=true&style=flat&logo=python)

  19
  
![requirements](https://img.shields.io/badge/Python-3.7-3480eb.svg?longCache=true&style=flat&logo=python)





  20
  

  20
  





  21
  
- Tensorflow-GPU==2.2.0 (tested on 2.1.0, 2.2.0, and 2.3.0) ```conda install tensorflow-gpu```

  21
  
- Tensorflow-GPU==2.2.0 (tested on 2.1.0, 2.2.0, and 2.3.0) ```conda install tensorflow-gpu```






  22
  
- TF-slim==1.1.0 ```pip install tf-slim==1.1.0```

  22
  
- TF-slim==1.1.0 ```pip install tf-slim==1.1.0```






  23
  
- Weights: [```download```](https://1drv.ms/u/s!AlXVRhh1rUKThlxTievX0X1CpXd0?e=9cKxYb) the trained weights file for

  23
  
- Weights: [```download```](https://1drv.ms/u/s!AlXVRhh1rUKThlxTievX0X1CpXd0?e=9cKxYb) the trained weights file for






  24
  
  blood cell detection and put the ```weights``` folder in the working directory.

  24
  
  blood cell detection and put the ```weights``` folder in the working directory.





  25
  

  25
  





  26
  
[![Download](https://img.shields.io/badge/download-weights-blue.svg?longCache=true&style=flat&logo=microsoft-onedrive)](https://1drv.ms/u/s!AlXVRhh1rUKThlxTievX0X1CpXd0?e=9cKxYb)

  26
  
[![Download](https://img.shields.io/badge/download-weights-blue.svg?longCache=true&style=flat&logo=microsoft-onedrive)](https://1drv.ms/u/s!AlXVRhh1rUKThlxTievX0X1CpXd0?e=9cKxYb)






  27
  
[![Download](https://img.shields.io/badge/download-weights-ff160a.svg?longCache=true&style=flat&logo=mega)](https://mega.nz/#F!2kVUnKjS!z15tM9WLfga3l1gCNSLNGw)

  27
  
[![Download](https://img.shields.io/badge/download-weights-ff160a.svg?longCache=true&style=flat&logo=mega)](https://mega.nz/#F!2kVUnKjS!z15tM9WLfga3l1gCNSLNGw)





  28
  

  28
  





  29
  
## Getting Started 

  29
  
## Getting Started 





  30
  

  30
  





  31
  
1. Build the cython extension in place 

  31
  
1. Build the cython extension in place 






  32
  
```python setup.py build_ext --inplace```

  32
  
```python setup.py build_ext --inplace```






  33
  
2. Run detect.py 

  33
  
2. Run detect.py 






  34
  
```python detect.py```

  34
  
```python detect.py```





  35
  

  35
  





  36
  
## Update

  36
  
## Update






  37
  
The ```darkflow.cython_utils.cy_yolo_findboxes``` problem has been fixed. Make sure to build the cython extension in place before running the code. 

  37
  
The ```darkflow.cython_utils.cy_yolo_findboxes``` problem has been fixed. Make sure to build the cython extension in place before running the code. 





  38
  

  38
  





  39
  
[![Paper](https://img.shields.io/badge/TensorFlow-2.x-f57418.svg?longCache=true&style=flat&logo=tensorflow)](https://www.tensorflow.org/install)

  39
  
[![Paper](https://img.shields.io/badge/TensorFlow-2.x-f57418.svg?longCache=true&style=flat&logo=tensorflow)](https://www.tensorflow.org/install)





  40
  

  40
  





  41
  
The code was originally written and developed with `TensorFlow v1.x`. The new updated version `v2.0`

  41
  
The code was originally written and developed with `TensorFlow v1.x`. The new updated version `v2.0`






  42
  
included `TensorFlow v2.x` support, tested on both TensorFlow `v2.1.0` and `v2.2.0`. You can download the previous

  42
  
included `TensorFlow v2.x` support, tested on both TensorFlow `v2.1.0` and `v2.2.0`. You can download the previous






  43
  
version

  43
  
version






  44
  
from **[`here`](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/releases/tag/v1.0)**

  44
  
from **[`here`](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/releases/tag/v1.0)**






  45
  
.

  45
  
.





  46
  

  46
  





  47
  
## How to Run the Code  :runner:

  47
  
## How to Run the Code  :runner:





  48
  

  48
  





  49
  
To detect the blood cells, simply run the `detect.py` file in the terminal or use an IDE. A step-by-step guideline of

  49
  
To detect the blood cells, simply run the `detect.py` file in the terminal or use an IDE. A step-by-step guideline of






  50
  
how to run the blood cell detection code in your computer is provided in

  50
  
how to run the blood cell detection code in your computer is provided in






  51
  
this **[`wiki`](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/wiki/A-Step-by-Step-Guide-of-How-to-Run-the-Code)**

  51
  
this **[`wiki`](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/wiki/A-Step-by-Step-Guide-of-How-to-Run-the-Code)**






  52
  
.

  52
  
.






  53
  
If you have any trouble running the code and facing any errors please feel free to create

  53
  
If you have any trouble running the code and facing any errors please feel free to create






  54
  
an **[`issue`](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/issues)**

  54
  
an **[`issue`](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/issues)**






  55
  
or **[`contact me`](https://mahmudulalam.github.io/#contact)**.

  55
  
or **[`contact me`](https://mahmudulalam.github.io/#contact)**.





  56
  

  56
  





  57
  
## How to Train on Your Dataset  :bullettrain_side:

  57
  
## How to Train on Your Dataset  :bullettrain_side:





  58
  

  58
  





  59
  
A seven-step guideline of how to train on your own dataset is provided in

  59
  
A seven-step guideline of how to train on your own dataset is provided in






  60
  
this **[`wiki`](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/wiki/How-to-Train-on-Your-Dataset)**

  60
  
this **[`wiki`](https://github.com/MahmudulAlam/Automatic-Identification-and-Counting-of-Blood-Cells/wiki/How-to-Train-on-Your-Dataset)**






  61
  
.

  61
  
.





  62
  

  62
  





  63
  
## Paper

  63
  
## Paper





  64
  

  64
  





  65
  
[![Paper](https://img.shields.io/badge/paper-IETDigiLib-830ceb.svg?longCache=true&style=flat)](http://ietdl.org/t/kmgztb) [![Paper](https://img.shields.io/badge/paper-Wiley-282829.svg?longCache=true&style=flat)](https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/htl.2018.5098)

  65
  
[![Paper](https://img.shields.io/badge/paper-IETDigiLib-830ceb.svg?longCache=true&style=flat)](http://ietdl.org/t/kmgztb) [![Paper](https://img.shields.io/badge/paper-Wiley-282829.svg?longCache=true&style=flat)](https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/htl.2018.5098)





  66
  

  66
  





  67
  
The code was developed for the following blood cell detection paper. For a more detailed explanation of the proposed

  67
  
The code was developed for the following blood cell detection paper. For a more detailed explanation of the proposed






  68
  
method, please go through the pdf of the [```paper```][1]. If you use this code or associated dataset, please cite this

  68
  
method, please go through the pdf of the [```paper```][1]. If you use this code or associated dataset, please cite this






  69
  
paper as:

  69
  
paper as:





  70
  

  70
  





  71
  
[***```Machine learning approach of automatic identification and counting of blood cells```***](https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/htl.2018.5098)

  71
  
[***```Machine learning approach of automatic identification and counting of blood cells```***](https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/htl.2018.5098)





  72
  

  72
  





  73
  
```bibtex

  73
  
```bibtex






  74
  
@article{alam2019machine,

  74
  
@article{alam2019machine,






  75
  
  title={Machine learning approach of automatic identification and counting of blood cells},

  75
  
  title={Machine learning approach of automatic identification and counting of blood cells},






  76
  
  author={Alam, Mohammad Mahmudul and Islam, Mohammad Tariqul},

  76
  
  author={Alam, Mohammad Mahmudul and Islam, Mohammad Tariqul},






  77
  
  journal={Healthcare Technology Letters},

  77
  
  journal={Healthcare Technology Letters},






  78
  
  volume={6},

  78
  
  volume={6},






  79
  
  number={4},

  79
  
  number={4},






  80
  
  pages={103--108},

  80
  
  pages={103--108},






  81
  
  year={2019},

  81
  
  year={2019},






  82
  
  publisher={IET}

  82
  
  publisher={IET}






  83
  
}

  83
  
}






  84
  
```

  84
  
```





  85
  

  85
  





  86
  
## Blood Cell Detection Output

  86
  
## Blood Cell Detection Output





  87
  

  87
  





  88
  
<p align="center">

  88
  
<p align="center">






  89
  
  <img src="https://user-images.githubusercontent.com/37298971/44617785-17eb0980-a88b-11e8-9018-c84f8be5cefa.png" width="500">

  89
  
  <img src="https://user-images.githubusercontent.com/37298971/44617785-17eb0980-a88b-11e8-9018-c84f8be5cefa.png" width="500">






  90
  
</p>

  90
  
</p>





  91
  

  91
  





  92
  
## KNN and IOU Based Verification

  92
  
## KNN and IOU Based Verification





  93
  

  93
  





  94
  
In some cases, our model predicts the same platelet twice. To solve this problem we propose a k-nearest neighbor (KNN)

  94
  
In some cases, our model predicts the same platelet twice. To solve this problem we propose a k-nearest neighbor (KNN)






  95
  
and intersection over union (IOU) based verification system where we find the nearest platelet of a selected platelet

  95
  
and intersection over union (IOU) based verification system where we find the nearest platelet of a selected platelet






  96
  
and calculate their overlap. We are allowing only a 10% overlap between two platelets. If the overlap is more than that

  96
  
and calculate their overlap. We are allowing only a 10% overlap between two platelets. If the overlap is more than that






  97
  
then it will be a spurious prediction and we will ignore the prediction.

  97
  
then it will be a spurious prediction and we will ignore the prediction.





  98
  

  98
  





  99
  
| Before Verification  | After Verification  |

  99
  
| Before Verification  | After Verification  |






  100
  
|:-:|:-:|

  100
  
|:-:|:-:|






  101
  
| <p align="center"> <img src="https://user-images.githubusercontent.com/37298971/46260207-b27ede00-c504-11e8-9d00-7d7be151ee5d.jpg"> </p>  | <p align="center"> <img src="https://user-images.githubusercontent.com/37298971/46260504-a268fd80-c508-11e8-9ef0-5230d00f47a3.jpg">  |

  101
  
| <p align="center"> <img src="https://user-images.githubusercontent.com/37298971/46260207-b27ede00-c504-11e8-9d00-7d7be151ee5d.jpg"> </p>  | <p align="center"> <img src="https://user-images.githubusercontent.com/37298971/46260504-a268fd80-c508-11e8-9ef0-5230d00f47a3.jpg">  |





  102
  

  102
  





  103
  
## Prediction on High-Resolution Image (HRI)

  103
  
## Prediction on High-Resolution Image (HRI)





  104
  

  104
  





  105
  
We have used our model to detect and count blood cells from high-resolution blood cell smear images. These test images

  105
  
We have used our model to detect and count blood cells from high-resolution blood cell smear images. These test images






  106
  
are of the size of ```3872 x 2592``` way higher than the size of our trained images of ```640 x 480```. So, to match the

  106
  
are of the size of ```3872 x 2592``` way higher than the size of our trained images of ```640 x 480```. So, to match the






  107
  
cell size of our trained images we divide those images into grid cells and run prediction in each grid cell and then

  107
  
cell size of our trained images we divide those images into grid cells and run prediction in each grid cell and then






  108
  
combine all the prediction results.

  108
  
combine all the prediction results.





  109
  

  109
  





  110
  
<h3 align="center">Dividing Image into Grid/Patch</h3>

  110
  
<h3 align="center">Dividing Image into Grid/Patch</h3>






  111
  
<p align="center">

  111
  
<p align="center">






  112
  
  <img src="https://user-images.githubusercontent.com/37298971/45962420-a39ab600-c042-11e8-975f-9b0a077f0e0f.jpg" width="700">

  112
  
  <img src="https://user-images.githubusercontent.com/37298971/45962420-a39ab600-c042-11e8-975f-9b0a077f0e0f.jpg" width="700">






  113
  
</p>

  113
  
</p>





  114
  

  114
  





  115
  
<h3 align="center">Combined Output</h3>

  115
  
<h3 align="center">Combined Output</h3>






  116
  
<p align="center">

  116
  
<p align="center">






  117
  
  <img src="https://user-images.githubusercontent.com/37298971/45961699-055a2080-c041-11e8-95b0-1c8ac3c8875b.jpg" width="700">

  117
  
  <img src="https://user-images.githubusercontent.com/37298971/45961699-055a2080-c041-11e8-95b0-1c8ac3c8875b.jpg" width="700">






  118
  
</p>

  118
  
</p>





  119
  

  119
  





  120
  
[1]: https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/htl.2018.5098

  120
  
[1]: https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/htl.2018.5098