search-api — AI-Powered Hybrid Search Engine for Patient Location

AI-driven search engine for locating missing patients in healthcare institutions using hybrid facial recognition and textual analysis.

Report Bug

🇬🇧 English · 🇪🇸 Español

---

About search-api

IAPEX (Inteligencia Artificial para la Localización de Pacientes Extraviados en Instituciones de Salud) is a validated and robust system designed to address a critical problem: identifying unlocated or unidentified patients within healthcare institutions.

This repository contains the Search API — the AI engine that combines facial recognition via CNN and advanced textual analysis to identify patients with high precision.

Ecosystem

Component	Repository	Stack
Search API (this)	iapex-org/search-api	FastAPI, Python 3.11, MongoDB
Mobile App	iapex-org/mobile-app	React, Ionic
Web Portal	iapex-org/web-app	Angular, Bootstrap
Core API	iapex-org/core-api	Spring Boot, PostgreSQL

Features

• Facial Recognition — CNN based on dlib with 128-dimensional embeddings
• Textual Analysis — Levenshtein and Jaro-Winkler algorithms for matches
• Hybrid Search — combines facial recognition (75%) and textual analysis (25%)
• Smart Cache — face encodings in MongoDB for fast comparisons
• Ad-hoc Search — searches with manually entered data
• Validated System — 100% sensitivity (zero false negatives) in 3,720 evaluations
• Async API — async/await architecture for maximum performance
• Interactive Docs — Swagger UI and ReDoc built-in

Search Modes

Mode	Description	Validated Precision
text	Text-only (name, gender, physical description)	96.8% effectiveness
image	Facial recognition only via CNN	100% precision
hybrid	Optimal combination: text (25%) + image (75%)	100% precision

Quick Start

Prerequisites

• Python 3.11+
• PostgreSQL 14+
• MongoDB 6.0+
• CMake and C++ compilers (for dlib)

Setup

git clone https://github.com/iapex-org/search-api.git
cd search-api
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
cp .env.example .env

Configure your values in .env, then:

python run.py

The API will be available at http://localhost:8000

Architecture

┌─────────────────────────────────────────────────┐
│                  Client Apps                     │
│  ┌─────────────┐  ┌─────────────────────────┐   │
│  │ Mobile App  │  │      Web Portal         │   │
│  └──────┬──────┘  └───────────┬─────────────┘   │
└─────────┼─────────────────────┼─────────────────┘
          │                     │
┌─────────▼─────────────────────▼─────────────────┐
│           API SEARCH (FastAPI)                    │
│  ┌──────────┐ ┌──────────┐ ┌────────────────┐   │
│  │ Facial   │ │  Text    │ │   Hybrid       │   │
│  │ Recog.   │ │  Matcher │ │   Engine       │   │
│  │ (dlib)   │ │(Levensh.)│ │  (75/25 mix)   │   │
│  └────┬─────┘ └────┬─────┘ └───────┬────────┘   │
│       │            │               │            │
│  ┌────▼────────────▼───────────────▼────────┐   │
│  │         Cache Layer (MongoDB)            │   │
│  └────────────────┬─────────────────────────┘   │
└───────────────────┼─────────────────────────────┘
                    │
          ┌─────────▼─────────┐
          │  PostgreSQL       │
          │  (Patient Data)   │
          └───────────────────┘

Contributing

Please read CONTRIBUTING.md for branch naming, commit conventions, and PR workflow.

License

This project is licensed under the GNU General Public License v3.0 — see the LICENSE file for details.

Acknowledgments

Authors:

• Altamirano Misrael Florentino
• Ortiz Pérez Alejandro
• Serrano Puertos Jorge Christian

Advisors:

• Escobar García Arturo

Academic Support:

• Universidad Tecnológica del Centro de Veracruz