This project is detect the orientation of a malabar seed. Classify the orientation into up, down, side based on the location of the hilum. See full demo video here.
arduino_demo.py Demo with arduino and the whole mechanism . Computer: run arduino_demo.py Arduino: run ./../arduino/motor_control/motor_control.ino
real_time_demo.py Demo with webcam (laptop's webcam)
test.py test accuracy of Vgg19 with testing data.
demo_video Contain two videos. Each video records a real-time bud detecting process.
net Contain a python module named vgg19. Also, vgg19.py demos how to predict the label of an image
weights Contains three different weights of VGG19 Net.
- fine_tune_weight.npy Always load it when predicting labels
- unfine_tune_weight.npy Not important.
- vgg19.npy Pretrained weights which is pretrained on ILSVRC. Now important either.
data Contain training data and testing data. Literaly bunch of images.
train Contain requiring tools to train VGG19 Net. Don't execute it unless one knows how to train VGG19 Net.