Machine Perception is an automated intelligence tool that allows users to search and filter large video and image datasets for specific objects, anomalies, similar images, or 3D features. It aims to reduce annotation and labeling costs by 10x by providing a tool to wrangle and search through computer vision datasets.

To use Machine Perception, simply upload your large dataset of images or videos. You can then use the search and filter tool to narrow down your dataset based on text, similar images, or 3D features. The tool will provide you with the filtered results, allowing you to focus on the images that require annotation and saving you annotation and labeling costs.