Monocular 3D pose estimation. OpenVINO. CPU inference or iGPU (OpenCL) inference.

Last update: Oct 03, 2022

Overview

human-pose-estimation-3d-python-cpp

RealSenseD435 (RGB) 480x640 + CPU Corei9 45 FPS (Depth is not used)

1. Run

1-1. RealSenseD435 (RGB) 480x640 + CPU Corei9 45 FPS (Depth is not used)

$ xhost +local: && \
docker run -it --rm \
-v `pwd`:/home/user/workdir \
-v /tmp/.X11-unix/:/tmp/.X11-unix:rw \
--device /dev/video0:/dev/video0:mwr \
--device /dev/video1:/dev/video1:mwr \
--device /dev/video2:/dev/video2:mwr \
--device /dev/video3:/dev/video3:mwr \
--device /dev/video4:/dev/video4:mwr \
--device /dev/video5:/dev/video5:mwr \
--net=host \
-e XDG_RUNTIME_DIR=$XDG_RUNTIME_DIR \
-e DISPLAY=$DISPLAY \
--privileged \
ghcr.io/pinto0309/openvino2tensorflow:latest

$ python3 human_pose_estimation_3d_demo.py \
--model models/openvino/FP16/human-pose-estimation-3d-0001_bgr_480x640.xml \
--device CPU \
--input 4

1-2. RealSenseD435 (RGB) 480x640 + iGPU (OpenCL)

$ xhost +local: && \
docker run -it --rm \
-v `pwd`:/home/user/workdir \
-v /tmp/.X11-unix/:/tmp/.X11-unix:rw \
--device /dev/video0:/dev/video0:mwr \
--device /dev/video1:/dev/video1:mwr \
--device /dev/video2:/dev/video2:mwr \
--device /dev/video3:/dev/video3:mwr \
--device /dev/video4:/dev/video4:mwr \
--device /dev/video5:/dev/video5:mwr \
--net=host \
-e LIBVA_DRIVER_NAME=iHD \
-e XDG_RUNTIME_DIR=$XDG_RUNTIME_DIR \
-e DISPLAY=$DISPLAY \
--privileged \
ghcr.io/pinto0309/openvino2tensorflow:latest

$ python3 human_pose_estimation_3d_demo.py \
--model models/openvino/FP16/human-pose-estimation-3d-0001_bgr_480x640.xml \
--device GPU \
--input 4

1-3. General USB Camera 480x640 + CPU

$ xhost +local: && \
docker run -it --rm \
-v `pwd`:/home/user/workdir \
-v /tmp/.X11-unix/:/tmp/.X11-unix:rw \
--device /dev/video0:/dev/video0:mwr \
--net=host \
-e XDG_RUNTIME_DIR=$XDG_RUNTIME_DIR \
-e DISPLAY=$DISPLAY \
--privileged \
ghcr.io/pinto0309/openvino2tensorflow:latest

$ python3 human_pose_estimation_3d_demo.py \
--model models/openvino/FP16/human-pose-estimation-3d-0001_bgr_480x640.xml \
--device CPU \
--input 0

2. Build

$ PYTHON_PREFIX=$(python3 -c "import sys; print(sys.prefix)") \
&& PYTHON_VERSION=$(python3 -c "import sys; print(f'{sys.version_info.major}.{sys.version_info.minor}')") \
&& PYTHON_INCLUDE_DIRS=${PYTHON_PREFIX}/include/python${PYTHON_VERSION}

$ NUMPY_INCLUDE_DIR=$(python3 -c "import numpy; print(numpy.get_include())")

$ mkdir -p pose_extractor/build && cd pose_extractor/build

$ cmake \
-DPYTHON_INCLUDE_DIRS=${PYTHON_INCLUDE_DIRS} \
-DNUMPY_INCLUDE_DIR=${NUMPY_INCLUDE_DIR} ..

$ make && cp pose_extractor.so ../.. && cd ../..

Monocular 3D pose estimation. OpenVINO. CPU inference or iGPU (OpenCL) inference.

Related tags

Overview

human-pose-estimation-3d-python-cpp

1. Run

1-1. RealSenseD435 (RGB) 480x640 + CPU Corei9 45 FPS (Depth is not used)

1-2. RealSenseD435 (RGB) 480x640 + iGPU (OpenCL)

1-3. General USB Camera 480x640 + CPU

2. Build

3. Reference

Owner

Katsuya Hyodo

ProMP: Proximal Meta-Policy Search

Telegram chatbot created with deep learning model (LSTM) and telebot library.

One implementation of the paper "DMRST: A Joint Framework for Document-Level Multilingual RST Discourse Segmentation and Parsing".

3D-Transformer: Molecular Representation with Transformer in 3D Space

Semi-supervised Implicit Scene Completion from Sparse LiDAR

Si Adek Keras is software VR dangerous object detection.

Source code for our paper "Do Not Trust Prediction Scores for Membership Inference Attacks"

A curated list of awesome Model-Based RL resources

Lexical Substitution Framework

Dynamic Neural Representational Decoders for High-Resolution Semantic Segmentation

Unsupervised Feature Loss (UFLoss) for High Fidelity Deep learning (DL)-based reconstruction

一个运行在 𝐞𝐥𝐞𝐜𝐕𝟐𝐏 或 𝐪𝐢𝐧𝐠𝐥𝐨𝐧𝐠 等定时面板的签到项目

SW components and demos for visual kinship recognition. An emphasis is put on the FIW dataset-- data loaders, benchmarks, results in summary.

🔥 Cannlytics-powered artificial intelligence 🤖

git《Pseudo-ISP: Learning Pseudo In-camera Signal Processing Pipeline from A Color Image Denoiser》(2021) GitHub: [fig5]

Focal and Global Knowledge Distillation for Detectors

Practical tutorials and labs for TensorFlow used by Nvidia, FFN, CNN, RNN, Kaggle, AE

Solution to the first stage Quiz of Hamoye internship: Introduction to Python for Machine Learning

CVPRW 2021: How to calibrate your event camera

This is an official implementation of CvT: Introducing Convolutions to Vision Transformers.