🤝 FusionNet

Touchless Palmprint and Finger Texture Recognition with Deep Learning Feature Fusion

Source code for the IEEE CIVEMSA 2019 paper
Touchless palmprint and finger texture recognition: A Deep Learning fusion approach

---

🧠 Overview

FusionNet is a MATLAB implementation of a touchless biometric recognition pipeline that combines two complementary hand traits acquired from the same palmar image:

• Palmprint
• Inner Finger Texture (IFT)

The method extracts multiple Regions of Interest (ROIs), trains the same deep learning topology on each biometric trait, and performs feature-level fusion to improve recognition performance without requiring additional acquisitions.

---

✨ Key Ideas

• 🖐️ Single touchless hand acquisition
• 🌴 Palmprint ROI extraction
• ☝️ Inner Finger Texture ROI extraction
• 🧬 Deep feature extraction with a PCANet-inspired architecture
• 🔗 Feature-level fusion across palm and finger texture representations
• 📊 Biometric evaluation for touchless and less-constrained recognition

---

📌 Pipeline

Touchless hand image
        │
        ▼
Database processing
        │
        ▼
Palmprint and IFT ROI extraction
        │
        ├───────────────┬───────────────┬───────────────┐
        ▼               ▼               ▼               ▼
  Palmprint ROI      IFT-1 ROI        IFT-2 ROI        IFT-n ROI
        │               │               │               │
        ▼               ▼               ▼               ▼
 Deep feature     Deep feature     Deep feature     Deep feature
 extraction       extraction       extraction       extraction
        │               │               │               │
        └───────────────┴───────────────┴───────────────┘
                        │
                        ▼
                Feature-level fusion
                        │
                        ▼
              Matching and evaluation

---

📁 Repository Structure

FusionNet/
│
├── main_FusionNet.m                    # Main script
├── README.md                           # Project documentation
├── LICENSE                             # GPL-3.0 license
│
├── (0) Common functions/               # Shared utility functions
├── (A) Process DB files/               # Dataset loading and preprocessing
├── (B) ROI extraction/                 # Palmprint and finger texture ROI extraction
├── (C) PCANet_featureFusion/           # Deep feature extraction and fusion routines
│
└── images/
    ├── outline.png                     # Pipeline illustration
    └── DB Fusion Palm-Knuckle (orig)/
        └── REST_hand_database/         # Expected REST dataset location

---

🚀 Getting Started

1. Clone the repository

git clone https://github.com/AngeloUNIMI/FusionNet.git
cd FusionNet

2. Prepare the REST hand database

Download the REST hand database from the official provider and place it in:

./images/DB Fusion Palm-Knuckle (orig)/REST_hand_database/

The expected folder structure is:

images/DB Fusion Palm-Knuckle (orig)/REST_hand_database/p1
images/DB Fusion Palm-Knuckle (orig)/REST_hand_database/p2
images/DB Fusion Palm-Knuckle (orig)/REST_hand_database/p3
...

Each pX folder should contain the corresponding hand images for that subject.

3. Run FusionNet

Open MATLAB, move to the repository folder, and run:

main_FusionNet

---

📊 Output

FusionNet performs the main stages required for touchless palmprint and finger texture fusion:

Stage	Description
Database processing	Reads and organizes REST hand images
ROI extraction	Extracts palmprint and Inner Finger Texture regions
Feature extraction	Computes deep features using the PCANet-inspired pipeline
Fusion	Combines palmprint and IFT information at feature level
Evaluation	Computes biometric recognition performance

---

🧪 Dataset

The experiments are based on the REST hand database:

Dataset	Link
REST Hand Database	http://www.regim.org/publications/databases/regim-sfax-tunisian-hand-database2016-rest2016/

---

🖥️ Demo Version

A demonstration version of FusionNet for webcam-based touchless palmprint and finger texture recognition is available here:

https://github.com/AngeloUNIMI/Demo_FusionNet

---

📚 Related Code and Dependencies

FusionNet includes or uses code inspired by the following works and libraries:

• T. Chan, K. Jia, S. Gao, J. Lu, Z. Zeng, and Y. Ma,
  “PCANet: A Simple Deep Learning Baseline for Image Classification?”
  IEEE Transactions on Image Processing, 2015.
  DOI: 10.1109/TIP.2015.2475625

• A. Vedaldi and B. Fulkerson,
  “VLFeat: An Open and Portable Library of Computer Vision Algorithms”, 2008.
  http://www.vlfeat.org/

• Peter Kovesi,
  MATLAB and Octave Functions for Computer Vision and Image Processing.
  https://www.peterkovesi.com/matlabfns/

---

📖 Paper

If you use this code, please cite:

@InProceedings{civemsa19,
  author    = {A. Genovese and V. Piuri and F. Scotti and S. Vishwakarma},
  title     = {Touchless palmprint and finger texture recognition: A Deep Learning fusion approach},
  booktitle = {Proc. of the 2019 IEEE Int. Conf. on Computational Intelligence and Virtual Environments for Measurement Systems and Applications (CIVEMSA 2019)},
  address   = {Tianjin, China},
  month     = {June},
  day       = {14--16},
  year      = {2019},
  pages     = {1--6},
  doi       = {10.1109/CIVEMSA45640.2019.9071620},
  isbn      = {978-1-5386-8344-6}
}

Paper:

https://ieeexplore.ieee.org/document/9071620

Project page:

http://iebil.di.unimi.it/fusionnet/index.htm

---

👥 Authors

• Angelo Genovese
• Vincenzo Piuri
• Fabio Scotti
• Sarvesh Vishwakarma

---

📄 License

This project is released under the GNU General Public License v3.0.

See the LICENSE file for details.