https://github.com/rosesonfire/objectdetection

Detect single objects in small, background-blurred and close-focused images
https://github.com/rosesonfire/objectdetection

api blurred-background close-focused image-editting image-processing nodejs object-detection

Last synced: 15 days ago
JSON representation

Detect single objects in small, background-blurred and close-focused images

• Host: GitHub
• URL: https://github.com/rosesonfire/objectdetection
• Owner: rosesonfire
• License: mit
• Created: 2017-11-19T09:27:36.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2017-11-22T14:10:03.000Z (about 7 years ago)
• Last Synced: 2024-11-28T12:07:28.736Z (24 days ago)
• Topics: api, blurred-background, close-focused, image-editting, image-processing, nodejs, object-detection
• Language: JavaScript
• Homepage:
• Size: 668 KB
• Stars: 6
• Watchers: 3
• Forks: 0
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
Detect single objects in small, background-blurred and close-focused images.



NPM link: [object-detection](https://www.npmjs.com/package/object-detection)

# Installation

`npm install object-detection`

# Usage

```

var detectObject = require('object-detection')

// Tweak the Sensitivity and the Tolerance for optimal result

var config = {

  imageName: 'path-to-image', // preferrably in JPG format and less than 100 kB

  sensitivity: 50, // ranges from 1 to 100

  tolerance: 50, // ranges from 1 to 100

}

detectObject(config).then(function(response) {

  var base64Img = response.base64Img

  // use base64Img in html image tag ...

  imageElement.setAttribute('src', 'data:image/jpeg;base64,' + base64Img)

  // save base64Img as image file ...

  fs.writeFile('object.jpg', base64Img, { encoding: 'base64' }, function() {

  	console.log('Saved object image')

  })

  

})

```

# Caution

- Use small images, preferably in JPG format and less than 100 kB

- Currently works for single object.

- Behaviour is not strongly defined for multi-objects.

- Contiguous objects are considered as single object.

- Optimizer is not well tuned yet.

# Web console

[object-detection-console](https://github.com/rosesonfire/objectDetectionConsole) is a simple web console for the object-detection package.

# Examples

Name|Image|Object|Sensitivity|Tolerance

-|-|-|-|-

Baseball|||87|50

Birb|||94|57

Cherry|||86|21

Mic|||87|20

Sunflower|||94|22

Flower|||80|17