Using Real-Time Object Tracking/Segmentation for AR project

There are 3 steps to this project, with 3 milestone payments.

We are creating an AR game using Unity as the game engine. We have been using Google ARCore but are unhappy with the drift and tracking capabilities for our needs.

We have seen many YOLO/YOLACT/Mask-RCNN tracking videos (like the link below) and are impressed with the real-time tracking it can produce.

[login to view URL]

For our app, a user will have a camera within a headset (POV), so for the solution, you can not used a fixed camera point.

Here are the 3 steps to the project.

Step 1: Proven ability to track a person's real-time location/direction within a full sized basketball court. This must be from a camera (POV).

Step 2: Proven ability to track a basketball and a basketball backboard in real-time from a camera (POV)

Step 3: Proven ability to highlight interactions between ball and backboard in real-time (for example, if both ball and backboard are being tracked, we need you to show a graphic element every time the ball touches the backboard).

PLEASE only apply if you are confident you can achieve thes 3 steps, otherwise you waste your time and mine.

Thank you.

Kemahiran: Tensorflow, Machine Learning (ML), Python, Pytorch, Deep Learning

Lihat lagi: cheap real time gps tracking, real time gps tracking blackberry hidden, real time gps tracking google maps mobile, samsung phone real time gps tracking software, real time gps tracking wifi, blackberry real time gps tracking, free real time gps tracking mobile phone, free real time gps tracking using gps phone, real time eye tracking using matlab, real time eye tracking project matlab, real time object detection opencv using surf, real time object tracking opencv python, real time object tracking, real time object detection using matlab, real-time object tracking github, real time object counting using raspberry pi, real-time object detection project report, real time object detection using yolov3, real time gps tracking project, real time object detection using webcam

ID Projek: #29421823