We are currently working on an exciting proprietary project which involves image placement and detection in virtual space using depth sensing with LiDAR and non-LiDAR enabled phones. We’ve an aggressive rollout schedule and looking for consultants with experience in computer vision who can help us resolve a particular problem.
Preferred devices: iPhone XS and iPhone 11
Preferred Development Platform/SDK: ARKit 4
Image Placement: Using a non-LiDAR enabled phone, a user should be able to place a virtual 2D image and video on a flat surface (interior and exterior) with minimal (ideally none) gap between the surface and the image.
Image Detection: Using a non-LiDAR enabled device, a user should be able to quickly detect a 2D image and video already placed in the virtual space. On a LiDAR enabled device, a user should be able to quickly detect a 2D image and video placed a distance of more than 5 meters.
The device should be able to detect a vertical and horizontal surface easily (In minimal time).
User should be able to place an image /video on a desired surface with minimal gap between the image and surface.
The placed image / video should be stable and provide a perspective view. For instance, the user should be able to move around the image as if it were a real one.
The placed image should allow for ocular view. For instance, if the image is placed on a wall and the user were to move his/her hand over the image, the image would be hidden by the movement of the hand.
For non-LiDAR devices (any range) and LiDAR devices (with range greater than 5m), the user should be able to detect 2D images / videos quickly (in minimal time).
The detected image should allow for ocular view.
1. Image should not float. It should be seen as close a real world image placed on a shape
2. Nice to have: if the digital images placed on a 3D real world object should conform to the shape of the object
3. Placing the image on a street view on a building, window, billboard etc.. with distance of >5m and a minimum 500 meters or one CDM proper what is the minimum viewable distance. When a user walks towards or away from the image. The image should increase in size or reduce in size and remain in the place holder e.g. Window or billboard in relation to the size of the window or billboard based on the users viewing distance. Ensuring that the image looks and feels as close to viewing the same with naked eye.
Image Placement & Detection: We have been able to achieve these with LiDAR devices within a 5m range.