Pikapi is a perception library for interaction with agents.
Project description
PIKAPI
This repository contains perception algorithms for interaction with agetns. The name comes from capitals of "Perception toward Interaction with Kawaii Agents and Pretty Individuals". My focus is making the world full of kawaii characters.
Requirements
For use perception algorithms
For python, please install following modules.
pip install pikapi
To build library from source.
This repository is currently tested on Ubuntu 18.04 and python3.7.7. So it might work around these OS and python versions.
- requirements to run mediapipe
- (Running facemesh example for testing successful installation is recommended.)
- protobuf-compiler
For python, please install following modules.
- numpy
- pyqt5
- scipy
- opencv > 3
Perception Algorithms
Head Gesture Recognition
This app is for recognizing head gesture. Currently supported gesture is nodding and shaking head. (Video)
If you install the module with pip
GLOG_minloglevel=2 python apps/recognize_head_gesture.py
If want to compile from source.
# To run on GPU.
python setup.py build_ext
GLOG_minloglevel=2 python apps/recognize_head_gesture.py
# To run on CPU.
cd modules/face
wget -O - https://raw.githubusercontent.com/PINTO0309/PINTO_model_zoo/master/030_BlazeFace/05_float16_quantization/download_new.sh | bash
wget -O - https://raw.githubusercontent.com/PINTO0309/PINTO_model_zoo/master/032_FaceMesh/05_float16_quantization/download.sh | bash
cd ../../
python3 tools/generate_proto.py
# If you use non-system python, it is necessary to point to the path of python.
# to make a python binding appropriately.
# export PYTHON_BIN_PATH=$YOUR_PYTHON_PATH
bash development/prepare_framework_bindings.sh
GLOG_minloglevel=2 python apps/recognize_head_gesture.py
Examples for mediapipe use in python
These are the example of getting the face landmark results from graph.
Python Graph Runner for GPU
Graph Runner API is an API with higher level than calculator framework API in mediapipe. This API is provided with python binding using pybind11 and it support GPU.
To run simple app using this graph runner API, please follow these steps.
python setup.py build_ext
python apps/run_face_mesh_live.py
Python Graph Runner for CPU
To run graph runner for cpu (pure python). Run the following command.
cd modules/face
wget -O - https://raw.githubusercontent.com/PINTO0309/PINTO_model_zoo/master/030_BlazeFace/05_float16_quantization/download_new.sh | bash
wget -O - https://raw.githubusercontent.com/PINTO0309/PINTO_model_zoo/master/032_FaceMesh/05_float16_quantization/download.sh | bash
cd ../../
python3 tools/generate_proto.py
# If you use non-system python, it is necessary to point to the path of python.
# to make a python binding appropriately.
# export PYTHON_BIN_PATH=$YOUR_PYTHON_PATH
bash development/prepare_framework_bindings.sh
python apps/run_face_mesh_live_cpu.py
Development Code
Code that is used to develop algorithms are included under /development
.
License
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
File details
Details for the file pikapi-0.0.1.dev1-cp37-cp37m-manylinux2014_x86_64.whl
.
File metadata
- Download URL: pikapi-0.0.1.dev1-cp37-cp37m-manylinux2014_x86_64.whl
- Upload date:
- Size: 23.3 MB
- Tags: CPython 3.7m
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.2.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/46.1.3.post20200330 requests-toolbelt/0.9.1 tqdm/4.45.0 CPython/3.7.7
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 171bb643a200853a58cd157dd8e496e4170d8ddd3f3dee9b0cc130bba9a4771f |
|
MD5 | aa7ace6616e9926ec37180f64cb16898 |
|
BLAKE2b-256 | b8a495af8f5fc61dc079251992c89bd1481a37f00ba23f491792df60fe0fc3f0 |