Mostafa-Hesham1/Vehicle-Souq

AI-powered platform for Egypt's used car market featuring vehicle recognition (94.41% accuracy), price prediction (97% R²), and damage detection using DenseNet201, XGBoost, and YOLOv8. Published at IEEE IMSA 2025 Conference. # 🚗 Vehicle Souq ### AI-Powered Car Recognition, Price Prediction & Damage Detection System Click the image above to watch our full demo video --- ## 📌 Overview Vehicle Souq is a comprehensive AI-powered platform designed to revolutionize Egypt's used car market. The system addresses critical challenges including price inconsistencies, lack of transparency, and difficulty in assessing vehicle condition. ### Key Capabilities | Feature | Description | Accuracy | |---------|-------------|----------| | 🔍 **Vehicle Recognition** | Identify make, model, year, and body type from a single image | 94.41% | | 💰 **Price Prediction** | Fair market valuations based on 23,421+ real listings | 97% (R²=0.97) | | 🔧 **Damage Detection** | AI-powered assessment using YOLOv8/Mask R-CNN | mAP@50: 0.87 | | 📊 **Market Analytics** | Data-driven insights for buyers, sellers, and analysts | Real-time | ### The Challenge Egypt's used car market faces unique challenges: - **Price Volatility** - Inflation and market instability create unpredictable pricing - **Information Asymmetry** - Buyers lack expertise to evaluate vehicles accurately - **Manual Valuation** - Traditional methods are time-consuming and inconsistent - **Condition Assessment** - Difficulty in identifying and quantifying vehicle damage ### Our Solution Vehicle Souq uniquely combines three AI systems: 1. **DenseNet201** for vehicle recognition (94.41% accuracy) 2. **XGBoost** for price prediction (97% accuracy, R² = 0.97) 3. **YOLOv8/Mask R-CNN** for damage detection and segmentation --- ## 🎓 Research & Publication This project has been published and presented at the **2025 Intelligent Methods, Systems, and Applications (IMSA)** conference, held in Giza, Egypt. **📄 Used Car Price Prediction and Classification Using Machine Learning Approaches** **Authors:** M. Hesham, H. Ahmed, M. L. Borham **Institution:** MSA University, Egypt **DOI:** 10.1109/IMSA65733.2025.11167441 Certificate of Par …