alrikf / wastedata-mask_rcnn-multiple-classes

predicting more than 2 classes from publicly available image data, for annotation we are using VGG annotator latest version

wastedata-multiple-classes

This Repo is just a tiny modification of Matterports MaskRCNN Repo link is here, the reason why i didnt fork the original one is because re-uploading the necessary modified part is much easier and cleaner to go through.

Instance segmenting more than 2 classes in an image,where image dataset is from publicly available data, for annotation we are using VGG annotator latest version

Here we defined 4 classes :

• bottle
• glass
• paper
• trash

Below are the examples for model's accuracy

Bottle

Glass

Trash

Four in one

This is the collage of four classes in one picture, and the prediction result

Since we just tweaked a bit on original code of matter port's mask-rcnn, it do has all the step by step detection

Color splash

Anchor sorting and filtering

Bounding box

Mask Generation

# Train a new model starting from pre-trained COCO weights

python final.py train --dataset=/path/to/datasetfolder --weights=coco

# Resume training a model that you had trained earlier

python final.py train --dataset=/path/to/datasetfolder --weights=last