Classification of defects in semiconductor wafer using artificial intelligence
Machine learning, a subset of artificial intelligence is an emerging technology that enabled the classification of objects without the need of being explicitly programmed. Due to the popularity of artificial intelligence, many frameworks were invented. ANN, CNN, Faster RCNN will be explained t...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/78800 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Machine learning, a subset of artificial intelligence is an emerging technology that enabled
the classification of objects without the need of being explicitly programmed. Due to the
popularity of artificial intelligence, many frameworks were invented. ANN, CNN, Faster RCNN
will be explained to understand the fundamentals of machine learning. However, the
focus of this project is on the framework Mask-RCNN, developed by Facebook, it uses
region-based convolutional neural network that simultaneously perform object detection and
instance segmentation.
This project comprises of two important part. The first step is to obtain the datasets in the
form of images in large numbers of a 1000. The images are annotated by drawing polygons
on the region of interest and a json file is obtained. The Mask R-CNN is downloaded on the
computer and a virtual environment is created, dependencies are installed for training to take
place. The second part includes running the training to obtain the h5 files. Detection is run to
determine the success of the training. The whole process will be repeated if the detection is
unable to produce the results needed. |
|---|