Fuzzy Calculating of Human Brain’s Weight Using Depth Sensors

Link to the paper:

• https://www.researchgate.net/publication/328558987_Fuzzy_Calculating_of_Human_Brain's_Weight_Using_Depth_Sensors

Please cite :

• Mousavi, Seyed Muhammad Hossein. "Fuzzy Calculating of Human Brain’s Weight Using Depth Sensors." 2nd Iranian Symposium on Brain Mapping (ISBM2018). Vol. 10. National Brain Mapping Lab (NBML), 2018.

Abstract

This paper presents a novel, cost-effective, and non-invasive method for estimating human brain weight using depth sensors, such as Kinect V.2. The proposed method is aimed at detecting abnormalities like Microcephaly and Macrocephaly, which are typically diagnosed through expensive and time-consuming MRI imaging. By leveraging depth data and fuzzy classification, the method achieves promising results with a recognition accuracy of 97.2% for three classes: Normal, Microcephaly, and Macrocephaly.

Key Features:

• Fast and Low-Cost: Uses depth sensors instead of MRI, reducing time and financial costs.
• Non-Invasive: Does not expose individuals to magnetic fields or radiation.
• Real-Time Applicability: Designed for real-time processing and decision-making.
• Fuzzy Classification: Employs fuzzy logic for robust and interpretable classification.

---

Methodology

1. Data Acquisition:

   • Two depth images (front and side views) are recorded using Kinect V.2 sensors.
   • The depth data is preprocessed to create a 3D model of the subject's head.

2. 3D Model Reconstruction:

   • Depth images are registered using non-rigid registration techniques to create a full 3D representation of the head.
   • Skull and skin volumes are removed (~10% of the total volume) to estimate brain volume.

3. Brain Weight Calculation:

   • The brain volume is normalized based on age-specific average brain weights.
   • The volume is converted to weight using a custom mapping.

4. Fuzzy Classification:

   • The estimated brain weight is classified into one of three categories: Normal, Microcephaly, or Macrocephaly.
   • Fuzzy sets and linguistic variables are used for classification.

5. Validation:

   • Tested on 10 subjects (6 normal, 4 abnormal).
   • Achieved a mean squared error (MSE) of 2.8% compared to actual brain weights.

---

Results

• Recognition Accuracy: 97.2%
• Classification Categories:
  • Normal
  • Microcephaly
  • Macrocephaly
• Mean Squared Error (MSE): 2.8%

---