Official PyTorch implementation of "I2L-MeshNet: Image-to-Lixel Prediction Network for Accurate 3D Human Pose and Mesh Estimation from a Single RGB Image", ECCV 2020
MIT License
720
stars
128
forks
source link
What FPS could we expect for the inference time? #37
Have you done an estimation of the time it takes for the inference + rendering? Is it possible to use this model in real time applications, for example ~10 FPS? How much will the inference time increase if we have multiple people?
Let's say I am using YOLOv4 as a detection pipeline, so we can ignore the detection overhead.
I2L-MeshNet runs at 50fps for each person as written in the paper. This time does not include the rendering, but I think rendering will take marginal time.
Have you done an estimation of the time it takes for the inference + rendering? Is it possible to use this model in real time applications, for example ~10 FPS? How much will the inference time increase if we have multiple people?
Let's say I am using YOLOv4 as a detection pipeline, so we can ignore the detection overhead.