Segment the fields of the bankbook
Segment the fields of the bankbook
posted in 2019
posted in 2019
(I) Objective: Segment all the fields on the inner page of the bankbook.
(II) Approach:
Mask R-CNN with the ResNet50 feature extractor.
Data augmentation: scale, translate, rotate, additive Gaussian noise, multiply(multiply all pixels in an image with a specific value), gray scale.
(III) Results: The bankbook results of the Mask R-CNN with the ResNet50 :
(test case 1) :
(Fig. 1a) input image
(Fig. 1b) the predicted result with the bounding boxes
(test case 2) :
(Fig. 2a) input image
(Fig. 2b) the predicted result with the bounding boxes