-
Requirements for Tensorflow(1.0 and so far)
-
Python(3.5) packages : 'tflearn', 'opencv3', 'skimage', 'PIL(pillow)', 'matplotlib', 'numpy', 'sklearn'
-
ImageNet dataset
-
selectivesearch.py : selective search source code
-
selectivesearch_example.py : example of selectivesearch
-
train_alexnet.py : pre-train for cnn model(alexnet, you can replace with vgg)
-
alexnet_test.py : test for cnn model
-
preprocessing.py : some preprocess
-
rp2file.py : save region proposal images to disk
-
fine_tune.py : fine-tune for cnn model use region proposal images
-
train_svm.py : train svm
-
detection_softmax_test.py : object detection use softmax
-
detection_svm_test.py : object detection use svm
-
pre-train, $python train_alexnet.py
-
generate region proposal images, $python rp2file.py
-
fine-tune, $python fine_tune.py
-
train svm, $python train_svm.py
-
run detection, $python detection_softmax_test.py or detection_svm_test.py