Nazar

Electronic component detection system with server: Nazar Server

Overview

Nazar is an app built with react-native with a little spice of Tensorflow InveptionV3 which allows the user to take a picture using the camera or fetch image from gallery to identify the component with the predicted percentage ratio, thus using image processing and algorithm to segment them after detection along with fetching description about the detected component using Octopart API whose response is sent from the nazar-server itself.

It is done with both Clarifai API and Tensorflow frozen graph server deployed in heroku to deduce with internet along with option to look for feeds from internet within the app. The Option to fetch the details about detected component is setup but needs furnishing.

Run both on iOS and Android

Demo

Information

Installation

Clone the source locally:

$ git clone https://github.com/aryaminus/nazar
$ cd nazar

Start the application in development mode

npm install
react-native link
react-native run-android
react-native run-ios

or for VS-Code:

npm install
react-native link

then press F1 or Fn+F1 and React Native:Run Android on Device or React Native:Run iOS on Simulator

References

Contributing

Fork it (https://github.com/aryaminus/nazar/fork)
Create your feature branch (git checkout -b feature/fooBar)
Commit your changes (git commit -am 'Add some fooBar')
Push to the branch (git push origin feature/fooBar)
Create a new Pull Request

Undefined is not an object (tensorflow ImageRecognition)

When trying to integrate a pretrained tensorflow model with expo (react-native), the following error occurs within these lines:

_graph = async () => {
        try {
            const tfImageRecognition = new TfImageRecognition({
                model: require('./assets/tensorflow_inception_graph.pb'),
                labels: require('./assets/tensorflow_labels.txt')
            });
    
            const results = await tfImageRecognition.recognize({
              image: this.state.image
            }); 
            results.forEach(
              result => ((preder2 = result.confidence), (items = result.name))
            );
            await tfImageRecognition.close();  
            this.setState({
              result: items,
              value: preder2 * 100 + "%"
            });
            console.log(this.state.result);
          } catch (err) {
            this.setState({
              result: "No Internet",
              value: "Please connect to the internet"
            });
            console.log(err);
          }
    }

Which generates the following error

10:10:40 AM: undefined is not an object (evaluating 'RNImageRecognition.initImageRecognizer')

I have been trying to find the reason why this is not working but I cannot find a definite solution. The relative paths linking to the assets are correct and the extensions are present in the app.json. Furthermore the model is trained using the tensorflow api which should make it compatible with the react-native implementation.

I observed that after running

const tfImageRecognition = new TfImageRecognition({

                model: require('./assets/tensorflow_inception_graph.pb'),
                labels: require('./assets/tensorflow_labels.txt')

            });

The code changed immediately to "catch(err)" branch, which means it could not load the model and labels?

I am using expo SDK version 28.0.0, Expo XDE and react-native-tensorflow version ^0.1.8
PLz help me.

speks7 / nazar Goto Github PK

nazar's Introduction

Nazar

Overview

Demo

Information

Installation

References

Contributing

nazar's People

Contributors

Stargazers

Watchers

Forkers

nazar's Issues

Undefined is not an object (tensorflow ImageRecognition)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent