In this paper, we propose the simple method to optimize the datasets
noise under the uncertainty applied to many applications in industry.
Specifically, we use firstly the deep learning module at transfer
learning based on using the mask-rcnn to detect the objects and
segmentation effectively, then return the contours only. After that we
address the shortest path for reduce the noise in order to increasing
the highspeed in industrial applications. We illustrate adaptive many
applications web applications such as mobile application where power
computer is limited a source