Skip to content

ShijieZhou-UCLA/Video-Object-Segmentation

Repository files navigation

Video Object Segmentation

COMS W4995: Applied Deep Learning

Demos:

1. Reproduce Part (OSVOS):

parent_network_demo.ipynb: train parent network

test_network_demo.ipynb: train test network and test the results

2. Innovation Part (OSVOS + RCF):

RCF_parent_network_demo.ipynb: train parent network

RCF_test_network_demo.ipynb: train test network and test the results

Codes

Data loading:

data_parent.py: load data for parent network

data_finetune.py: load data for test network

Neural Network

cnn.py: functions and architecture of OSVOS network

cnn_rcf.py: functions and architecture of OSVOS + RCF network

Train:

train_parent.py: train function for parent network

train_finetune.py: train function for test network

train_parent.txt: list of training data

Test:

test_finetune.py: test function for test network

Others

DAVIS: dataset and result images

models: training models of reproduce part*

models_RCF: training models of innovation part*

*: Due to the 100mb file restriction of Github, the files in these two folders cannot be uploaded. Please download here and replace them.

About

COMS W4995 Applied Deep Learning

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published