Hello I have a question regading to dpt

#4
by Fulanxi - opened

Well i'll start simply. All ai tracking characters has problem to estimate the depth of the character properly. Then the bones could be out but if the character in moving from far to close to the camera or even running + plus the camera is moving, we got a bones system based only on what the Ai is ready, static picture....When we import the bones system generated, it looks static, bone here, moving almost correctly but not moving on 3d space...I was considering if we can use the dpt information each frame to help the AI to move the character in (XYZ) correctly? Camera can be tracked outside of course. Or maybe the AI need the camera tracked, can help even more?

Thank you

Sign up or log in to comment