Custom Vision Introduction

Custom Vision works through a simple process of uploading images, tagging them, and training a model. Users can bring their own labeled images or use the service to quickly add tags to unlabeled images. The system then uses these labeled images to train a custom vision model tailored to the specific use case. Once trained, the model can be tested and refined through an iterative process. Custom Vision offers both image classification (applying labels to whole images) and object detection (identifying and locating specific objects within images). The trained models can be accessed via API calls, allowing for easy integration into various applications and workflows. Models can also be exported for offline use on mobile devices or edge computing scenarios.