README.md

Toothbrush-Inspection

Detect defect bamboo toothbrush with CNN based algorithm

This project is implemented to achieve 4 module detection

1. Detect defect toothbrush from frontal toothbrush image
2. Detect defect crack from frontal toothbrush image
3. Detect defect side toothbrush from side toothbrush image
4. Detect defect crack from back toothbrush image

Environment

pip install -r requirements.txt

1. Download Datset

download dataset from datasets

download datasets.tar and untar

tar -xvf datasets.tar 

2. Demo

0) inference 4 modules in real time

python new_main.py

1) inference with trained model

download models from trained models

and place it ..

/models/back_crack/mask_rcnn_toothbrush_crack_0069.h5
/models/brush/mask_rcnn_toothbrush_head_0020.h5
/models/brush/efficient-best_weight_220119_2.h5
/models/brush/eff0_220928_2.h5
/models/front_crack/mask_rcnn_toothbrush_crack_0084.h5

1-1) inference with your trained model with custom data

place wherever you want

---

1. Detect defect toothbrush from frontal toothbrush image

python toothbrush_head_final.py

2. Detect defect crack from frontal toothbrush image

python toothbrush_crack_final.py

3. Detect defect side toothbrush from side toothbrush image

python toothbrush_side_final.py

4. Detect defect crack from back toothbrush image

python toothbrush_back_final.py

2) inference 4 modules in real time (faster .ver using multiprocess)

python multi_que.py

3. Visualize

1. Detect defect toothbrush from frontal toothbrush image

python toothbrush_head_final_visualize.py

example images :