officialarijit/glaucoma-classification-ml-dl

Glaucoma and Non-Glaucoma classification using ML/Dl and ensemble approaches using Image Feature Extraction Using HOG (Histogram of Gradient)

Glaucoma and Non-Glaucoma classification using ML/Dl and ensemble approaches using Image Feature Extraction Using HOG (Histogram of Gradient)

Packages

• Jupyter-Notebook : will be used to run the programs.

  pip install scikit-learn

• Open CV : For reading and manipulating images.

  pip install opencv-python

• Numpy : used for multi-dimensional arrays and matrices.

  pip install numpy

• Sklearn : will be used to get PCA for dimensionality reduction.

  pip install scikit-learn

• Mlxtend : will be used to implement the Majority Voting Ensemble approach.

  pip isntall mlxtend

Run The Program

• Run the program by executing the below code.

  run Feature Extraction HOG.ipynb

Result

• Result is generated and placed inside the

  100_extracted_features.csv.

Classification

• Once the features are extracted then run the following code. run Glaucoma-Ensemble-Approach.ipynb

  -- In this file Multi-Layer Perceptron (MLP), SVM and Random Forest (RF) classifiers are available. Also, the Majority Voting Ensemble approach is available.

glaucoma-classification-ml-dl's Issues

index out of range in feature extraction code

# Reshaped the Features to the acurrate size
reshaped_hog_Array = np.reshape(
    hogArray_np, (hogArray_np.shape[0], hogArray_np.shape[2]))

here the dimension of hogArray_np is NxN and 'hogArray_np.shape[2]' this will be out of range,isnt it?