Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/justsecret123/one-piece-image-classifier

A quick image classifier trained with manually selected One Piece images.
https://github.com/justsecret123/one-piece-image-classifier

batch-normalization classification classification-model computer-vision convolutional-neural-networks data-augmentation deep-learning deep-neural-networks docker early-stopping image-classification image-classifier jupyterlab keras machine-learning notebook python tensorflow tensorflow-serving transfer-learning

Last synced: 29 days ago
JSON representation

A quick image classifier trained with manually selected One Piece images.

Host: GitHub
URL: https://github.com/justsecret123/one-piece-image-classifier
Owner: Justsecret123
License: gpl-3.0
Created: 2021-01-05T16:19:25.000Z (about 4 years ago)
Default Branch: master
Last Pushed: 2022-05-01T06:42:39.000Z (over 2 years ago)
Last Synced: 2024-11-05T18:34:11.802Z (3 months ago)
Topics: batch-normalization, classification, classification-model, computer-vision, convolutional-neural-networks, data-augmentation, deep-learning, deep-neural-networks, docker, early-stopping, image-classification, image-classifier, jupyterlab, keras, machine-learning, notebook, python, tensorflow, tensorflow-serving, transfer-learning
Language: Jupyter Notebook
Homepage:
Size: 83.3 MB
Stars: 2
Watchers: 1
Forks: 1
Open Issues: 0
Metadata Files:
- Readme: README.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md

Awesome Lists containing this project

README

        # One-Piece-Image-Classifier ![Language_support](https://img.shields.io/pypi/pyversions/Tensorflow) ![Last_commit](https://img.shields.io/github/last-commit/JustSecret123/Human-pose-estimation) ![Workflow](https://img.shields.io/github/workflow/status/JustSecret123/Human-pose-estimation/Pylint/main) ![Tensorflow_version](https://img.shields.io/badge/Tensorflow%20version-2.6.2-orange)

I trained a custom Deep Learning model to recognize some One Piece characters (17, at the moment) with Tensorflow, Keras, and a fine-tuned MobilenetV2.

Afterwards, the model was converted to a TFLite model for running inferences on small devices, and a Tensorflow Serving Docker container for HTTP-based inferences.

96.84% validation accuracy.

> **Deployed on my personal Docker Hub repository: [Click here](https://hub.docker.com/repository/docker/ibrahimserouis/my-tensorflow-models)

> **Kaggle Notebook link:  [Kaggle notebook](https://www.kaggle.com/ibrahimserouis99/one-piece-image-classifier-notebook)

> **Tensorflow Lite model : [op_classifier_V16.tflite](https://drive.google.com/file/d/1-xTYqv10IQmV0FS3ftYH6LeTbE86dlkT/view?usp=sharing)



  



## Training 

A fine-tuned mobilenet has been used. The training session was runned on Kaggle with a GPU execution type. 

### Dataset 

> Dataset link : [Click here](https://www.kaggle.com/datasets/ibrahimserouis99/one-piece-image-classifier)

### Notebook 

> Notebook link : [Click here](one-piece-image-classifier-notebook.ipynb)

## Model description

- **A data augmentation layer** which creates "modified" images of the training set

- **A MobileNet layer** which detects the features

- **A Global average pooling layer** which converts the feature vectors into a 1280 element vector

- **1*3 Dense layers** followed by a dropout layer to prevent overfitting

- **An activation layer** (sigmoid) which represents the final output: Probability of input(X) belonging to each class

- Output classes (17 probabilities) : ['Ace', 'Akainu', 'Brook', 'Chopper', 'Crocodile', 'Franky', 'Jinbei', 'Kurohige', 'Law', 'Luffy', 'Mihawk', 'Nami', 'Robin', 'Sanji', 'Shanks', 'Usopp', 'Zoro']

## Architecture

![Model architecture](Screenshots/Model_summary_v16.PNG)

> **Best valdiation accuracy: 96.84%.**

## Results (Amazon SageMaker endpoint)

![Results](Screenshots/One%20Piece%20image%20classifier.gif)

# Useful scripts 

- [JFIFF to PNG converter](/Scripts/convert_jfif_to_png.py)

- [Image inverter](Scripts/invert_images.py)

## How to use 

### 1-First option: Using the [command line runner](Scripts/op_classifier.py)

#### Args 

![Args](Screenshots/command_line_args.PNG)

> **The image source can be a file path or a URL. Set the "mode" parameter to 'image' or 'url' accordingly.**

#### Results 

![Results](Screenshots/command_line_results.PNG)

### 2-Second option: Using the Tensorflow Serving image deployed [here](https://hub.docker.com/repository/docker/ibrahimserouis/my-tensorflow-models) (**TAG: OP_serving**)

Pull the Docker image with the OP_serving tag, then run inferences using the 8501 port. 

> A test script example is available [here](/Scripts/Prediction_OP_Model_Test.py)

## Prerequisites

- Python 3.7 or higher 

- IDE: Jupyter Lab/Kaggle Notebooks/Google Colab 

- Frameworks: Tensorflow 2.6 or higher and its dependencies

- Libraries : OpenCV, PIL, NumPy