yfy324 / nii_image_classify Goto Github PK

This repository is for my SSP project at NII. The topic of the project is to classify images by feature engineering, with the assistance of SIFT and BoF algorithms. The classifer in this experiment is xgboost.

Here are the libraries utilized in the project:
python3.6
numpy 1.19.3
opencv-python 3.4.2.16
opencv-contrib-python 3.4.2.16
scikit-learn 0.23.2
xgboost 1.5.0
imgaug  0.4.0
matplotlib 3.2.1
pandas 1.1.2
scipy 1.5.2


The demo_SIFT.py is just for testing some basic library in opencv.

The files in the directory 'binary_classification' is applied to classify dog ('n02085620' in the ImageNet) and cat ('n02123045' in the ImageNet). 'bow' method in opencv2 is used generate the visual dictionary.

The files in the directory 'multi_classification' is applied to classify peacock ('n01806143' in the ImageNet), elephant ('n02504013' in the ImageNet), airliner ('n02690373' in the ImageNet) and cat ('n02123045' in the ImageNet). 'bow' method in opencv2 is used generate the visual dictionary. In order to improve the performance of the classifier, the visual dictionary in this task is generated by bow method written by myself.