Risk of injury can intimidate individuals from learning new exercises and can discourage a healthy lifestyle. A promising solution to this problem is a smart posture-detection system that provides real-time feedback to prevent injuries. Valens is a computer-vision powered intelligent gym mirror that overlays posture-correcting feedback on a user’s reflection. Visual sensors track key points on the user’s body and feed data into pose-estimation algorithms to determine the optimal posture for an exercise. This design offsets hardware complexity and costs using AI and computer-vision models.

Many small farms lack fire detection systems in their barns as a result of not having affordable options that suit their needs. SafeFarm is a solution that combines the use of infrared and computer vision to detect fires the moment that they occur. It incorporates an RF message delivery to provide an audible alarm when an incident occurs. SafeFarm's advantages are that it provides an inexpensive, real-time fire detection system that gives this population the peace of mind that their livelihood is safe.

CIRKOPT is a program to assist in the design of transistor-level circuits and to allow higher-quality circuit designs with less effort. It automatically optimizes transistor sizes using an intelligent search algorithm to find a solution that minimizes power, delay, and area. Manual transistor sizing takes a great deal of time and expertise, therefore automating this will produce better performing circuits faster and allow more designer effort to be focused on topological and architectural changes.

Ichnaea is an affordable and comprehensive object tracking system, that will allow users to track specific objects, analogous to a sports broadcasting setup. It will consist of dual-cameras set on a swivel, that will feed information to the backend, where computer vision software will detect and track the object, as well as, issue commands to rotate the camera, keeping the object in focus. This project will deliver greater benefits over existing alternatives due to its low-cost, reliable object tracking, and detailed statistical output.