https://github.com/andrewtyw/auto-image-classifier

A one-stop solution for image classification: after defining the subjects that need to be classified in a JSON file, the system will automatically scrap images for each class from Bing and use them to train a model for image classification.
https://github.com/andrewtyw/auto-image-classifier

deep-learning-application flask python vue web-scraper

Last synced: 6 days ago
JSON representation

A one-stop solution for image classification: after defining the subjects that need to be classified in a JSON file, the system will automatically scrap images for each class from Bing and use them to train a model for image classification.

• Host: GitHub
• URL: https://github.com/andrewtyw/auto-image-classifier
• Owner: andrewtyw
• Created: 2022-01-11T07:11:40.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2022-11-22T03:10:40.000Z (almost 2 years ago)
• Last Synced: 2024-10-06T16:05:29.808Z (about 1 month ago)
• Topics: deep-learning-application, flask, python, vue, web-scraper
• Language: Python
• Homepage:
• Size: 7.77 MB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Auto-image-classifier

### Prototype demo👇

[prototype deployment is here](http://120.24.230.237:81/)

### what is Auto-image-classifier?

Actually, the concept is quite simple. First you need to define the categories that you want to make a classification in `srape_train/fish.json`. Then the Web Scraper will get photos of these categories from Bing. Then train a model for classification.

### Setup

- Scrape & trian model

```shell

# first define the classification subjects in `/scrape_train/fish.json`

# then scrape web images of the subjects

cd scrape_train/

python fetch.py --limit 10  # the number of images for each class

# train model

python train_main.py 

```

- Deployment

```shell

cd flask_fish_reco/

python myapp.py

```

### File structure

`scrape_train/`: scraping photos and training.

`flask_fish_reco/`: is the web deployment with Flask.

`neural_vue/`: user interfaces using Vue.

### Prototype looks like 👉