Raspberry Pi Projects

Raspberry pi4 Tensorflow object detection

Raspberry pi 4 TensorFlow Object Detection

On June 2019 Raspberry pi announce new version of raspberry pi board.
Now we have a new raspberry pi 4 model B 1GB So try to run TensorFlow object detection and then compare with Raspberry pi3B+ also.

Install TensorFlow on Raspberry pi4

Add some dependency

sudo pip3 install grpcio

sudo pip3 install h5py

Then Download tensorFlow
https://www.piwheels.org/simple/tensorflow/tensorflow-1.13.1-cp37-none-linux_armv7l.whl#sha256=25f4ff027beec1e568baf8e90a07bad59d354560533d6b37318b9efeb70beeb1

sudo pip3 install tensorflow-1.13.1-cp37-none-linux_armv7l.whl

Summary Software for test.

Rasbien 10 ( Buster )
Python 3.7.3
OpenCV 4.0.0
Tensorflow 1.13.1

Run Tensorflow Object Detection

Model SSDlite Mobilenet V2
Video MP4 768x432 12 fps

run on the same img os in same sd-card.

Raspberry pi 4 is 2.xx fps .

Raspberry pi 3 is less than 1 fps.

So, only Raspberry pi Board is not faster enough for tensorFlow object detection application.

Then next we use raspberry pi with intel neural compute stick for better performance.

Raspberry pi 4 vs Raspberry pi3 specification

Raspberry pi OpenVINO with Intel Movidius ( Neural Compute Stick )

Raspberry pi OpenVINO with Intel Movidius

( Neural Compute Stick )

What is OpenVINO?

The Intel® Distribution of OpenVINO™ toolkit quickly deploys applications and solutions that emulate human vision. Based on Convolutional Neural Networks (CNN), the toolkit extends computer vision (CV) workloads across Intel® hardware, maximizing performance. The Intel Distribution of OpenVINO toolkit includes the Intel® Deep Learning Deployment Toolkit (Intel® DLDT).

OpenVINO™ toolkit, short for Open Visual Inference and Neural network Optimization toolkit, provides developers with improved neural network performance on a variety of Intel® processors and helps them further unlock cost-effective, real-time vision applications.The toolkit enables deep learning inference and easy heterogeneous execution across multiple Intel® platforms (CPU, Intel® Processor Graphics)—providing implementations across cloud architectures to edge devices. This open source distribution provides flexibility and availability to the developer community to innovate deep learning and AI solutions.

OpenVINO™ toolkit contains:

• Deep Learning Deployment Toolkit

• Open Model Zoo

The Intel® Distribution of OpenVINO™ toolkit is also available with additional, proprietary support for Intel® FPGAs, Intel® Movidius™ Neural Compute Stick, Intel® Gaussian Mixture Model - Neural Network Accelerator (Intel® GMM-GNA) and provides optimized traditional computer vision libraries (OpenCV*, OpenVX*), and media encode/decode functions. To learn more and download this free commercial product, visit: https://software.intel.com/en-us/openvino-toolkit

What is Intel Movidius ( Neural Compute Stick )?

The Intel® Movidius™ Neural Compute Stick (NCS) is a tiny fanless deep learning device that you can use to learn AI programming at the edge. NCS is powered by the same low power high performance Intel Movidius Vision Processing Unit (VPU) that can be found in millions of smart security cameras, gesture controlled drones, industrial machine vision equipment, and more.

more detail

http://raspberrypi4u.blogspot.com/2018/10/raspberry-pi-movidius-neural-compute-stick.html

Samples Code Demo

Face Detection , Object Detection ( Object Detection C++ Sample SSD )

Identify faces for a variety of uses, such as observing if passengers are in a vehicle or counting indoor pedestrian traffic. Combine it with a person detector to identify who is coming and going.

Pre-trained Face Detection model

https://download.01.org/openvinotoolkit/2018_R4/open_model_zoo/face-detection-adas-0001/FP16/face-detection-adas-0001.bin

https://download.01.org/openvinotoolkit/2018_R4/open_model_zoo/face-detection-adas-0001/FP16/face-detection-adas-0001.xml

To validate OpenCV* installation, you may try to run OpenCV's deep learning module with Inference Engine backend. Here is a Python* sample, which works with Face Detection model.

Face Detect Python Code on Github

Interactive Face Detection C++ Demo

Age & Gender Recognition

This neural network-based model provides age and gender estimates with enough accuracy to help you focus your marketing efforts.

Emotion Recognition

Identify neutral, happy, sad, surprised, and angry emotions.

Head Position Estimation

This model shows the position of the head and provides guidance on what caught the subject's attention.

This demo showcases Object Detection task applied for face recognition using sequence of neural networks. Async API can improve overall frame-rate of the application, because rather than wait for inference to complete, the application can continue operating on the host while accelerator is busy. This demo executes four parallel infer requests for the Age/Gender Recognition, Head Pose Estimation, Emotions Recognition, and Facial Landmarks Detection networks that run simultaneously. You can use a set of the following pre-trained models with the demo:

face-detection-adas-0001, which is a primary detection network for finding faces
age-gender-recognition-retail-0013, which is executed on top of the results of the first model and reports estimated age and gender for each detected face
head-pose-estimation-adas-0001, which is executed on top of the results of the first model and reports estimated head pose in Tait-Bryan angles
emotions-recognition-retail-0003, which is executed on top of the results of the first model and reports an emotion for each detected face
facial-landmarks-35-adas-0002, which is executed on top of the results of the first model and reports normed coordinates of estimated facial landmarks

For more information about the pre-trained models, refer to the https://github.com/opencv/open_model_zoo/blob/master/intel_models/index.md "Open Model Zoo" repository on GitHub*.

Security Barrier Camera С++ Demo

This code sample showcases vehicle detection, vehicle attributes, and license plate recognition.

The demo uses OpenCV to display the resulting frame with detections rendered as bounding boxes and text.

This demo showcases Vehicle and License Plate Detection network followed by the Vehicle Attributes Recognition and License Plate Recognition networks applied on top of the detection results. You can use a set of the following pre-trained models with the demo:

vehicle-license-plate-detection-barrier-0106, which is a primary detection network to find the vehicles and license plates
vehicle-attributes-recognition-barrier-0039, which is executed on top of the results from the first network and reports general vehicle attributes, for example, vehicle type (car/van/bus/track) and color
license-plate-recognition-barrier-0001, which is executed on top of the results from the first network and reports a string per recognized license plate

For more information about the pre-trained models, refer to the https://github.com/opencv/open_model_zoo/blob/master/intel_models/index.md "Open Model Zoo" repository on GitHub*.