🧠 DHARINI Full Lifespan 3D Fetal Brain Volume Navigator

An interactive, cloud-native 3D neuroimaging navigator designed to stream, reconstruct, and slice through high-resolution second-trimester human brain reconstructions from the IIT Madras DHARINI Dataset.

🙏 Acknowledgements & Gratitude

• To IIT Madras & The Sudha Gopalakrishnan Brain Centre: Deepest gratitude is extended to the research engineers, neuroanatomists, and acquisition teams at IIT Madras. Their painstaking work in digitizing, aligning, and publishing these massive cellular-resolution datasets pushes the boundaries of global computational medicine.
• To Sudha Murty & Philanthropic Visionaries: Sincere thanks to Sudha Murty and the philanthropic foundations whose foundational support and vision for cutting-edge computing centers in India have made world-class, open-access neuroscience research facilities a reality. Their generosity empowers independent developers worldwide to interact with high-value scientific pipelines.

Built independently by Srinivasta using Python, Streamlit, and GitHub, this system connects anonymously to the AWS Open Data Registry to execute real-time multi-planar slice extraction with a zero-RAM server footprint.

---

🚀 Live Application

🔗 Launch the Interactive Dashboard Web App

---

⚖️ Data Attribution & Project Disclaimer

• Independent Development: This software application is an independent open-source dashboard designed and architected entirely by Srinivasta. It is not financially sponsored by, affiliated with, or endorsed by IIT Madras or the Sudha Gopalakrishnan Brain Centre.
• Data Source Credit: The underlying 3D brain volumes utilized by this dashboard are sourced directly from the publicly accessible DHARINI Fetal Brain Atlas repository hosted on the AWS Registry of Open Data. Deep gratitude is extended to the original acquisition scientists for making their structural neuroimaging data open to the global research community.

---

🛠️ Key Architectural & Technical Capabilities

• On-Demand Voxel Streaming: Eliminates the need to host or download uncompressed multi-gigabyte files locally. The application leverages an anonymous, unsigned boto3 configuration to download raw dataset layers directly from the cloud on demand.
• Disk-Level Memory-Mapping (mmap='r'): To bypass Streamlit Cloud’s standard system RAM ceiling (<1GB), compressed .gz archives are unpacked directly into local NVMe disk blocks (/tmp/). nibabel then memory-maps the data structure, keeping the live runtime footprint under 50 Megabytes.
• Cinematic Fly-Through Engine: Implements a frame-by-frame animation pipeline using Streamlit session state hooks (st.session_state), enabling a smooth, looping fly-through visualization across 891 sequential tissue slices.
• Anatomical Contrast Boosting: Features a custom pixel-intensity masking and normalization workflow (NumPy-based background suppression) that isolates brain matter and dynamically stretches the grayscale dynamic range to highlight cellular migration layers.
• Institutional PDF Export Report: Integrates an on-demand clinical dossier generator (reportlab) that captures the active high-contrast image matrix from the browser view and compiles it into a static, publication-ready PDF document complete with automated milestone metadata logs.

---

📊 Dataset Matrix Coverage (AWS Open Data)

The application dynamically routes data for all 5 official specimens from the open-source repository:

1. Specimen 1 (14 Weeks - FB34): Early second-trimester baseline. Smooth cortical surfaces and primitive ventricular chambers.
2. Specimen 2 (17 Weeks - FB40): Initiation of rapid deep cellular migration. Bundled axon pathways around core zones.
3. Specimen 3 (21 Weeks - FB62): Mid-development stratification. Visual layering of the cortical plate and subplate zones.
4. Specimen 4 (22 Weeks - FB74): Structural expansion of the cerebral cortex.
5. Specimen 5 (24 Weeks - FB85): Late second-trimester complexity. Initial formation of physical sulci (folds) and highly defined cerebellum layout.

---

💻 Tech Stack & Dependencies

• Frontend UI Framework: Streamlit (with custom HTML/CSS responsive flexbox injection)
• Medical Data Operations: Nibabel (NIfTI-1 format parsing)
• Mathematical Compute Engine: NumPy (Array scaling, masking, matrix rotations)
• Cloud Storage Integration: Boto3 & Botocore (AWS S3 Unsigned SDK)
• Document Compilation: ReportLab (Vectorized PDF layout automation)

---

📥 How to Run Locally

1. Clone the Codebase

git clone https://github.com
cd dharini-brain-viewer

2. Install Project Requirements

Make sure you have Python 3.10+ active, then run:

pip install -r requirements.txt

3. Launch the Server App

streamlit run app.py

Your local browser tab will open automatically at http://localhost:8501.

---

📁 Repository Structure

├── app.py              # Main application logic & user interface layout
├── requirements.txt    # Project runtime dependency checklist
└── README.md           # Professional portfolio presentation documentation

---

🔬 Scientific Impact

This application lowers the barrier to entry for studying human brain development. By creating an interactive, web-accessible "digital twin," neuroscientists, medical students, and AI researchers can analyze critical volumetric landmarks—such as ventricular compression, gyrification timelines, and tissue density gradients—directly from any device without requiring high-end graphics processing hardware.

---

Developed and architected by Srinivasta as an open-source educational dashboard showcasing cloud data pipe optimization for high-throughput healthcare informatics.