A tool to use Custom Vision Prediction
Once you are on customvision.ai, create a project.
After your project is created, you will be on the tab training images but you don’t have once so, you will need to add them by clicking add images. Upload your pictures and after it, you will need to tag them. You can make a rectangle near what you want to be tagged and give it a tag.
After adding you pictures, you will need to train your model to recognize the tag on your pictures. You will need to click on the green button nammed Train. There you can choose if you prefer to make a quick training or an advanced training. An Advanced training will take more time but will be more precise in detection.
When the training is over, you can go in performance tab and find your iteration. If you want to be able to predict by our tool or by another app using the customvision API, you need to publish the iteration. You will need to choose which prediction ressource you will use to publish it. After this, you can go inside prediction URL and take the url prediction and the prediction key to put them later inside our tool.
You need to put your endpoint and key of customvision inside the settings part and after this, just click on the save button. After the save button clicked, you will be redirect inside the scanner part. There you can use the camera of your computer to detect things from your model.