Detection of Defects on Silicon Wafers

The task is to detect silicon wafers and classify them as “Good", “Small Defect” or “Big Defect”. This is a detection task followed by a classification task, which can be realized as a pipeline composed of a detector and a classifier.

Dataset:

The dataset contains 70 original images that mostly contain defective silicon wafers and 64 augmented images that only contain good silicon wafers. There are 3 types of silicon wafers, all annotated in LabelMe format as depicted below.

Data Loading & Annotation:

You have 2 options:

Download data annotated with LabelMe format. Load the imageset and the annotations that you downloaded, using ReliUI. To do that, import the imageset using “Import Data” and the annotation file per image folder using the “Import Annotated Dataset” (with the LabelMe annotation format option). Your annotated dataset will be available in the gallery.
Load the raw imageset and annotate the data yourself in any industry standard format you like using the ReliUI’s intuitive annotation functions. To do that, download the image folder. Follow the ReliVision Knowledge Hub User Guide (ReliUI: Data Curation) to define the target state/label (which is the label of the silicon wafer as good, small defect or big defect) to annotate the defects (using rectangle annotation tool as the task is object detection) and to save them in an appropriate industry standard format (for which LabelMe is a good option). A sample annotation of an image, where good, small defect and big defect silicon wafers are shown in green, red and yellow boxes, is as follows:

Building A Wafer Detection and Classification Pipeline:

We will be building a pipeline that is composed of a series of connected wafer detector and a wafer classifier. The wafer detector will be responsible for detecting the wafers, if there is any, and outputting a bounding box. The classifier will be fed with wafer detection results (the cropped images of the characters detected) and will decide on the type of the wafer (good, small defect or big defect).

Training a Wafer Detector:

1. Prepare your data for wafer detector training: In order to train a wafer detection model, you need to have training images in which all defects, of all types, are annotated with a common label, such as “wafer”. Merge your labels into a single class called “Wafer”.
2. Follow the ReliVision Knowledge Hub User Guide (ReliTrainer: Training an AI Block) to train your AI model for wafer detection. The main steps, detailed in the User Guide, include;

Model type selection: Detection
Annotated dataset selection
Automated or manual train/test split which essentially spares some data for training validation purposes.
Hyper parameter setting: We have chosen the following in this use case:

Epochs: 100 (default)
X-Y Resolution: 640 (default)
Learning Rate: 0.01 (default)
Momentum: 0.937 (default)
Weight Decay: 0.005 (default)

Sample training and validation loss and accuracy curves as a function of epochs are as follows:

Training a Wafer Classifier:

1. Prepare your data for wafer classifier training: In order to train a wafer classifier model, you need to have training images of individual (cropped) wafer cells with labels. The dataset provided has rectangle annotations with 3 labels corresponding to different label types. The best option is to
‍
Extract the rectangle annotations using the ReliVision data curation functions. See ReliVision Knowledge Hub User Guide to see how to do this.
‍
2. Follow the ReliVision Knowledge Hub User Guide (ReliTrainer: Training an AI Block) to train your AI model for whole image classification (good, small defect or big defect). The main steps, detailed in the User Guide, include;

Model type selection: Classification
Annotated dataset selection
Automated or manual train/test split which essentially spares some data for training validation purposes.
Hyper parameter setting: We have chosen the following in this use case:

Epochs: 100 (default)
X-Y Resolution: 224 (default)
Learning Rate: 0.0001 (default)
Momentum: 0.99 (default)
Weight Decay: 0.0005 (default)

Sample training and validation loss and accuracy curves as a function of epochs are as follows:

Building and Running the Pipeline:

Use the pipeline editor to build your pipeline by dragging and dropping the AI/Basic blocks. You will need to select an input data source (your raw image set), a detection AI Block (the defect detector you trained) and a defect classifier (the classifier you trained) connected in series as depicted below. Follow the ReliVision Knowledge Hub User Guide (ReliTrainer: Models / AI Pipeline).

Simply run your pipeline using the execute button at the top. You can review your results using the ReliUI data annotation interface. Your pipeline’s outputs will be saved as a separate set of annotations for each image it is run on. Visit the ReliVision Knowledge Hub User Guide (ReliTrainer: Pipeline Editor / AI Pipeline) for more details. Here is an output example:

Try it by yourself.
Download ReliVision now!

AI Model Train / Test

AI powered solution design-train/test-build functions of the ReliVision are performed in the ReliVision's Training Server [ReliTrainer] which is accessible through the user-friendly ReliUI application. You can train detection, segmentation and classification models, test your models, build complex visual analytics pipelines using your trained AI blocks (AI models) together with Basic blocks (Digital signal/image processing - DSP/DIP - functions).

Refer to ReliTrainer user guide for further details. AI Model Train/Test functions are available with a licensed ReliTrainer.

Deployment

AI powered visual analytics solutions that are built, trained and tested, can be deployed to ReliAudit through the user friendly ReliUI interface. The deployed solutions are run, managed and monitored at the edge by the ReliAudit. The user feedback received from the ReliAudit can also be pulled back to ReliTrainer via ReliUI for retraining.

Refer to ReliAudit user guide for further details. Deployment functions are available with a licensed ReliAudit.