title	IsekA.I
emoji	💻
colorFrom	gray
colorTo	gray
sdk	docker
app_port	7860
pinned	false
thumbnail	https://cdn-uploads.huggingface.co/production/uploads/6a4a628f07ad1b9980b1f953/D2jfq2djBJNIbq2yYfK9t.jpeg
short_description	Detect if anime art is Human-made or AI-generated

isekA.I — AI vs Human Anime Illustration Detector

🎓 Tugas Akhir / UAS Mata Kuliah: Perilaku Manajerial

A modern AI-powered web application that classifies anime illustrations as either Human-made or AI-generated using a fine-tuned EfficientNetB0 deep learning model.

Proyek ini dikembangkan sebagai bentuk pemenuhan Tugas Akhir / Ujian Akhir Semester (UAS) pada mata kuliah Perilaku Manajerial.

---

✨ Features

• 🎨 Premium Modern UI — Clean, minimal design inspired by modern AI SaaS products.
• 🧠 EfficientNetB0 Model — Fine-tuned on 6,485 anime illustrations for accurate classification.
• 📤 Drag & Drop Upload — Seamless image upload with instant preview.
• ⚡ Real-time Prediction — Fast inference with confidence visualization.
• 📊 Animated Progress Bars — Visual probability distribution of classification.
• 🌙 Dark Mode — Toggle between light and dark themes smoothly.
• 📱 Responsive Design — Works seamlessly on desktop, tablet, and mobile devices.
• 📋 Prediction History — Track recent classification results within the active session.
• 💾 Download Results — Export prediction details as a text document.
• 📋 Copy to Clipboard — Quick-copy prediction metrics.

---

🛠️ Tech Stack

Component	Technology
Backend	Flask (Python 3)
ML Model	TensorFlow / Keras
Architecture	EfficientNetB0
Frontend	HTML5, CSS3 (Vanilla), JavaScript (Vanilla)
Typography	Google Fonts — Poppins & Outfit
Icons	Font Awesome 6

---

📁 Project Structure

isekA.I/
├── app.py                              # Flask application & routes
├── utils.py                            # Image preprocessing & model inference
├── requirements.txt                    # Python dependencies
├── README.md                           # Project documentation (this file)
├── best_EfficientNet_finetuned.keras   # Pre-trained deep learning model
│
├── templates/
│   └── index.html                      # Main HTML page structure
│
├── static/
│   ├── css/
│   │   └── style.css                   # Custom styles & dark mode layout
│   ├── js/
│   │   └── script.js                   # Client-side validation & AJAX handling
│   ├── uploads/                        # Auto-created directory for temporary uploads
│   └── img/                            # Static images
│
└── assets/                             # Documentation assets

---

🚀 Quick Start

Prerequisites

• Python 3.8+
• pip (Python package installer)

Installation

1. Navigate to the project directory:

   cd isekA.I

2. Create and activate a virtual environment (optional but recommended):

   • Windows:

     python -m venv .venv
     .venv\Scripts\activate

   • macOS/Linux:

     python3 -m venv .venv
     source .venv/bin/activate

3. Install the dependencies:

   pip install -r requirements.txt

4. Run the Flask development server:

   python app.py

5. Access the application: Open http://127.0.0.1:5000 in your web browser.

---

🧠 Model Details

Property	Value
Architecture	EfficientNetB0
Training Dataset	6,485 images
Original Dataset	6,553 images
Classes	Human Illustration, AI Generated
Input Size	224 × 224
Transfer Learning	ImageNet
Fine-Tuning	Enabled
Test Accuracy	82.19%

Preprocessing Pipeline

1. Open Image → Convert to RGB color mode.
2. Resize → Downscale/scale to 224 × 224 px.
3. Array Conversion → Convert image to float32 NumPy array.
4. Dimension Expansion → Expand to batch dimension (1, 224, 224, 3).
5. EfficientNet Preprocessing → Scale pixel values using EfficientNet defaults.
6. Inference → Evaluate through fine-tuned Keras model (Sigmoid activation output).
7. Decision Boundary → Threshold score at 0.5 (Score $\ge$ 0.5 is AI Generated, < 0.5 is Human Art).

---

📡 API Reference

GET /

Renders the main dashboard page.

POST /predict

Processes image files and returns classification statistics.

• Content-Type: multipart/form-data
• Field: file (Image file upload)

Response Example:

{
    "error": false,
    "prediction": "AI Generated",
    "emoji": "🤖",
    "confidence": 96.84,
    "ai_probability": 96.84,
    "human_probability": 3.16,
    "processing_time": 0.245,
    "raw_score": 0.9684,
    "image_url": "/static/uploads/filename.png"
}

---

📝 Credits & Course Context

• Mata Kuliah: Perilaku Manajerial
• Status Proyek: Tugas Akhir / Ujian Akhir Semester (UAS)