Gather as many images as you can for a particular object in different resolutions, camera angles, lighting conditions, etc. and put them into a folder.
Bound and label each image to train the model to identify the object in each image. You can also auto-bound and then adjust accordingly.
Package and send the dataset to the cloud for training. Once, complete, the model can be uploaded to the Vaidio AI Vision Platform for use.
Use as you wish within Vaidio. You can always further train the AI model by using it to train itself on metadata images from its deployment.