Wahaha1314 / Fish-characteristic-measurement

Combined with mask R-CNN, each part of the fish is segmented and feature extraction is performed. (It contains a complete set of training samples)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

In order to better manage my project, the project has been transferred to https://github.com/YuChuang1205/Fish-characteristic-measurement-code. If you have any questions, please go to https://github.com/YuChuang1205/Fish-characteristic-measurement-code.



In the "Complete code" folder:
"train_model.py" is the file that performs the training.
"test_model.py" is the file that executes the test.
"train_data" folder is a training sample set.
"test_result.py" is the folder where the segmentation result graph is saved.
"logs" is the folder that generates the trained model.

In the "Total data set" folder:
"Test set" is the folder of the test set, which contains test sets in a pure background and in a complex background.
"Training samples" is the folder for the training sample set. It contains training sets for each part of the fish (which contains the annotation files required for training.)

When training, change the name of the corresponding study object in "train_model.py".
When testing, change the name of the corresponding study object in "test_model.py".

About

Combined with mask R-CNN, each part of the fish is segmented and feature extraction is performed. (It contains a complete set of training samples)


Languages

Language:Jupyter Notebook 99.5%Language:Python 0.5%