Anyone has an interesting project to build with Kinect? I have a 360 kinect laying around, since nobody does Just Dance parties anymore. Also, was the XBone version of kinect much better than 360?
I worked on a few projects where the Kinect was mounted on a self-driving robot for 3D mapping of areas unaccessible to humans.
Also worked in using it in applications with people with Alzheimer's.
The issue is very often lack of or outdated SDKs for the target platform. The hardware could be better but it's already getting many project to 90%.
An example: https://www.youtube.com/watch?v=GDZoOnzLYGo
V1 was based on an Israeli structured light tech that later became Apple TrueDepth, a static dot pattern taken with a camera which deviations becomes depth. V2 was technically an early and alternative implementation of Time-of-Flight LIDAR based on a light phase delaying device, that output deviations from set known distance as pictures or something along that.
There wouldn't have been lacking app support issues if V2 did work. There was/were something that made users be "yeah... by the way" about it.
The problem was that it ONLY had a windows SDK and most of the people who did amazing work outside of games and the xbox with the kinect v1 were using it with macs and linux in tools like openframeworks and processing. The v1 was developed outside microsoft and primesense and there were SDKs that were compatible cross platform. Tons of artists dove right in.
The KinectV2 only offered a windows sdk, and that's what killed it in the 'secondary market'.
Luckily now we have libfreenect2 (https://github.com/OpenKinect/libfreenect2) but that came too late for most of us.
Did MS provide SDKs for other OSes for v1 or was it just easier for people to make an open source one?
The 360 Kinect can only track two skeletons (but differentiate 6).
In order to get a good result, someone would need to train a good model for HPE that could use the cloud point data directly, but it seems nobody cares about depth sensor anymore, most efforts are going to HPE from regular 2d video (like media pipe holistic model). And given the result you can get with media pipe, openpose and the likes, it's understandable nobody is bothering working with low resolution cloud point anymore for 3D HPE.
The only use-case I can think of for a Kinect v1 in 2025, would be robotics if you want a low latency low resolution cloud point for your robot control, but even there I think we are moving to big vision model capable of making sense of regular video feeds.