Viola jones algorithm python

Saved searches

Use saved searches to filter your results more quickly

You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session.

Python implementation of the face detection algorithm described by Paul Viola and Michael J. Jones in 2001

salvacarrion/viola-jones

This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Sign In Required

Please sign in to use Codespaces.

Launching GitHub Desktop

If nothing happens, download GitHub Desktop and try again.

Launching GitHub Desktop

If nothing happens, download GitHub Desktop and try again.

Launching Xcode

If nothing happens, download Xcode and try again.

Launching Visual Studio Code

Your codespace will open once ready.

There was a problem preparing your codespace, please try again.

Latest commit

Git stats

Files

Failed to load latest commit information.

README.md

Python implementation of the face detection algorithm describe in Rapid Object Detection using a Boosted Cascade of Simple Features by Paul Viola and Michael J. Jones.

1. Selecting Haar-like features
2. Creating an integral image
3. Running AdaBoost training
4. Creating classifier cascades

Animation of the Haar-like features:

About

Python implementation of the face detection algorithm described by Paul Viola and Michael J. Jones in 2001

Источник

Saved searches

Use saved searches to filter your results more quickly

You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session.

Viola-Jones Adaboost Algorithms

hjt486/ViolaJonesAdaboost

This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Sign In Required

Please sign in to use Codespaces.

Launching GitHub Desktop

If nothing happens, download GitHub Desktop and try again.

Launching GitHub Desktop

If nothing happens, download GitHub Desktop and try again.

Launching Xcode

If nothing happens, download Xcode and try again.

Launching Visual Studio Code

Your codespace will open once ready.

There was a problem preparing your codespace, please try again.

Latest commit

Git stats

Files

Failed to load latest commit information.

readme.md

Viola-Jones Algorithm with Adaboost using Python 3

This is a Python 3 implementation of Viola-Jones Algorithm, please follow the instructions below to train the detector and using it to detect real photos.

• pickle # Used for saving calculated features, array-ized dataset onto disk.
• opencv # Used for reading and processing image files and convert them into a single array.
• scikit-learn # Used for accelerating the feature selection, you can comment them out in ‘ViolaJonesFaceDetector.py’ but expect significant slow process on feature selection.
• matplotlib.pyplot # Used for plotting box of features on example photo.

1. Download/clone the entire repository, all training/testing photos are stored in ‘dataset’ folder, you can change the dataset or using your own version of dataset.
2. If you’re using the dataset provided, you should be able to run ‘RunTrainerAndTester.py’ directly, but check to make sure the dependency of packages are met. You can open the file to change some parameters in TrainDetector(1, ‘new’, faces_images_count, non_faces_images_count, window_size)
   1. The first argument is an integer of how many Adaboost rounds you want;
   2. ‘new’ means generating new features, this is to save your time if your dataset is fixed (as features are calculated based on training dataset), so for the first time, using ‘new’ to generate features, and once the features are generated (saved in the same folder as ‘features.pkl’, about 500 MB for given dataset), then later on you can using the existed features to calculate different rounds of Adaboost classifiers, or change the parameters.
   3. Don’t change the rest of parameters unless you understand the program.

   by default, the maximum window size for Haar’s feature is 8 pixels to limit the training process.

   499 faces images and 2000 non-faces images are loaded and saved in Training-Data.pkl Training Started. Building features. Total features for 19 x 19 image with window size of 8 is 26168 Applying features to Training set. Calculating features on data # 0 / 2499 Calculating features on data # 1 / 2499 Calculating features on data # 2 / 2499 . Calculating features on data # 2496 / 2499 Calculating features on data # 2497 / 2499 Calculating features on data # 2498 / 2499 Start training weak classifiers selecting best classifier out of 2617 24.99% finished, please wait. 49.98% finished, please wait. . 74.97% finished, please wait. 99.96% finished, please wait. -------------Round # 10 ----------------- Alpha: 1.1e+01 Feature: twoVerticle ( 10 , 5 ) Width 2 Height 8 Training Error: 0.00% Training Treshold: -4.0 Training Accuray: 55.37% ( 526 / 950 ) Training False Positive: 42.11% ( 400 / 950 ) Training False Negative: 2.53% ( 24 / 950 ) --------------------------------------- success: Training Done --------------------------------------- Test dataset loaded 472 faces images and 475 non-faces images are loaded and saved in Test-Data.pkl [(('twoVerticle', 17, 4, 2, 6), -2.0, (769, 950)), (('twoVerticle', 3, 1, 3, 2), -2.0, (541, 950)), (('twoVerticle', 2, 1, 3, 2), -2.0, (404, 950)), (('twoVerticle', 4, 3, 2, 2), 0.0, (649, 950)), (('twoVerticle', 1, 1, 2, 8), -4.0, (530, 950)), (('twoVerticle', 3, 14, 3, 2), -1.0, (517, 950)), (('twoVerticle', 4, 7, 2, 8), 7.0, (480, 950)), (('twoVerticle', 4, 18, 3, 2), -1.0, (581, 950)), (('twoVerticle', 1, 2, 1, 2), -1.0, (481, 950)), (('twoVerticle', 11, 6, 2, 8), -4.0, (526, 950))] [5.079995271174973, 5.806817285138741, 8.16708732838109, 8.80540572350739, 6.696829303810116, 10.818388812842256, 7.827263950074671, 11.92781628687352, 10.144053913574604, 10.98478973594049] getBoxImage ('twoVerticle', 4, 18, 3, 2) ---------------Test Result------------ Adaboost 10 Round(s) Top Feature: twoVerticle ( 3 , 17 ) Width 3 Height 2 Top Threshold: -1.0 Top Training Accuracy: 61.16% ( 581 / 950 ) Test Accuracy: 66.10% ( 626 / 947 ) Test False Positive: 4.33% ( 41 / 947 ) Test False Negative: 29.57% ( 280 / 947 )

   For the last testing round, you may expected a plot pops up to show you the best feature (having largest alpha) as:

   

   please close the window to continue print out the test result.

   How to use real-world photo for face detection:

   Run ‘RunFaceDetectorOnRealPhoto.py’, or change the path yourself by openning the file and change ‘testimage’. It may take a while to display the result as below:

   Image testimage.jpg is loaded Start recognition.

   Источник

   Saved searches

   Use saved searches to filter your results more quickly

   You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session.

   A face detection program in python using Viola-Jones algorithm.

   Donny-Hikari/Viola-Jones

   This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.

   Name already in use

   A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

   Sign In Required

   Please sign in to use Codespaces.

   Launching GitHub Desktop

   If nothing happens, download GitHub Desktop and try again.

   Launching GitHub Desktop

   If nothing happens, download GitHub Desktop and try again.

   Launching Xcode

   If nothing happens, download Xcode and try again.

   Launching Visual Studio Code

   Your codespace will open once ready.

   There was a problem preparing your codespace, please try again.

   Latest commit

   Git stats

   Files

   Failed to load latest commit information.

   README.md

   Viola-Jones Detection Framework

   This is an implement of Viola-Jones Detection Framework and is used in human face detection.

   This implement of Viola-Jones Framework require python version 3.5.2, and depends on the following modules:

   module	version	comment
   numpy	1.13.3
   scipy	1.0.0
   opencv-python	3.4.0.14	for capturing image
   scikit-learn	0.19.1	for shuffling data

   Run the following command to begin face detection:

   Haar-like features are proposed by Viola and Jones, adapting the idea (from Papageoriou et al) of using Haar wavelets.

   

   In this implement, five types of Haar-like features are used. They are: left-right, top-bottom, horizontal-middle, vertical-middle, diagonal.

   

   To speed up features extraction process, an intermediate representation for the image called integral image is used.

   

   AdaBoost is short for Adaptive Boosting, which is a kind of method of ensemble learning.

   For more info about AdaBoost, refers to GitHub — Donny-Hikari/AdaBoost

   To speed up detection process, as well as to decrease the false positive rate while sustaining a high detection rate (i.e. to improve precision), Viola and Jones invent the boosted cascade. That is, to cascade multiple AdaBoost classifiers.

   

   1. Yoav Freund; Robert E. Schapire. AT&T Labs – Research, Shannon Laboratory. Journal of Japanese Society for Artificial Intelligence,14(5):771-780, September 1999. A Short Introduction to Boosting.
   2. Paul Viola; Michael J. Jones. IEEE Computer Society Conference on Computer Vision and Pattern Recognition, May 2004. Rapid Object Detection Using a Boosted Cascade of Simple Features.
   3. Paul Viola; Michael J. Jones. International Journal of Computer Vision 57(2), 137–154, 2004. Robust Real-Time Face Detection.

   About

   A face detection program in python using Viola-Jones algorithm.

   Источник