PINTO0309

OpenCV Text Detection (EAST text detector) 320x320 + EdgeTPU

Jetson Nano + GPU (OpenGL ES3.2) + MediaPipe + HandPose + Python 3.6

Jetson Nano + GPU (OpenGL ES3.2) + MediaPipe + Selfie Segmentation + Python 3.6

Jetson Nano + GPU (OpenGL ES3.2) + MediaPipe + FaceMesh + Python 3.6

Jetson Nano + GPU (OpenGL ES3.2) + MediaPipe + Pose + Python 3.6

CenterNet MobileNetV2 FPN Keypoints 320x320 Float32 + TFLite 20 Threads

RaspberryPi4 CPU only + RaspberryPi OS 64bit + Python3.7 + MediaPipe Hand Detection, 9 FPS

OnnxGraphQt - Practice ONNX editing operations with GUI

[25-30FPS] RaspberryPi4 CPU only + Tensorflow Lite + BlazeFace (Integer Quantization) + 4 Threads

Jetson Nano + GPU (OpenGL ES3.2) + MediaPipe + FaceDetection + Python 3.6

Facial Cartoonization, 60FPS, OpenVINO FP16, Corei7 CPU only, without GPU

DBFace (640x480) + OpenVINO(FP32) + USB Cam (640x480) + Corei7 only

EDN-GTM Dehazing, ONNX + TensorRT Execution Provider 384x640 Float16

MediaPipe Objectron (3D Object Detection, Chair) Weight Quant, Tensorflow Lite, x86 CPU 4Threads

RealSense D435+Depth+RaspberryPi3(Raspbian Stretch)

OpenVINO + Open3D + BlazePose 3D + USB Camera

MODNet - WebCam-Based Portrait Video Matting Demo 512x672

FaceMesh + Ubuntu 18.04 + GLES

Edge TPU + Posenet + Python

[4.3 FPS] MobileNetV2 base Openpose + WebCam + Tensorflow-CPU (disabled OpenVINO/Tensorflow Lite)

Real_Time_Image_Animation

TensorSpace.js - YoloV2

3D Human Pose Estimation + Tensorflow.js + Browser + WebGL

Head Pose Estimation - OpenCV

NanoDet "Super fast and lightweight anchor-free object detection model. Real-time on mobile devices"

Generate saved_model,tfjs,tf-trt,TPU,CoreML,Qauntized tflite,ONNX,OpenVINO,blob and pb from .tflite.

LattePanda Alpha Core m3 + USB 3.0 + Edge TPU Accelerator + MobileNet-SSD v2 + Async, 320x240

[45 FPS] Multi-TPU DeeplabV3, TPUx3 boosted

tiny-YoloV3 + NCS2 x1 + Accuracy tuning [20% ＞ 40%]

1 year GitHub commit history for PINTO0309 - 3D STL

[11-12 FPS/Core i7 CPU only] OpenVINO+DeeplabV3+Core i7-8750H RealTime semantic-segmentaion [Part.4]

[30 FPS++] MobileNetV2 base Openpose + WebCam + Tensorflow-GPU + GTX1070

OpenVINO + Mobile DeeplabV3-plus (MobileNetV3) FP16 + NCS2 (MYRIAD)