ygjwd12345 / Maskrcnn-hand

use the model from thematterport/Mask_RCNN and train the dataset from the website http://vision.soic.indiana.edu/projects/egohands/

Maskrcnn-hand

use the model from the matterport/Mask_RCNN and train the dataset from the website http://vision.soic.indiana.edu/projects/egohands/

Color Splash Example

This is an example showing the use of Mask RCNN in a real application. We train the model to detect hands only, and then we use the generated masks to keep hands in color while changing the rest of the image to grayscale.

Installation

From the Releases page page:

1.Download mask_rcnn_coco.h5 Save it in the root directory of the repo .

2.Download egohands_data.zip and egohands_video.zip Expand it such that it's in the path hand/.

Run Jupyter notebooks

Open the inspect_hand_data.ipynb or inspect_hand_model.ipynb Jupter notebooks. You can use these notebooks to explore the dataset and run through the detection pipelie step by step.

Train the Hand model

Train a new model starting from pre-trained COCO weights

python3 hand.py train --dataset=/path/to/hand/dataset --weights=coco

Resume training a model that you had trained earlier

python3 hand.py train --dataset=/path/to/hand/dataset --weights=last

Train a new model starting from ImageNet weights

python3 hand.py train --dataset=/path/to/hand/dataset --weights=imagenet

The code in hand.py is set to train for 3K steps (30 epochs of 100 steps each), and using a batch size of 2. Update the schedule to fit your needs.

video

video