Deep Learning Based Compressive Domain Analytics Framework for Seismic Images

Authors

Siddharath Narayan Shakya, Parimala Kancharla
ICVGIP '24: Proceedings of the Fifteenth Indian Conference on Computer Vision Graphics and Image Processing
Article No.: 40, Pages: 1 - 9
DOI: https://doi.org/10.1145/3702250.3702290

---

📄 Paper

Read the full paper here: ACM Digital Library

---

🎥 Video Presentation

Watch the video summary on YouTube: Deep Learning Based Compressive Domain Analytics Framework

---

🔬 Checkpoints

Salt Traces Identification in Compressive Domain

Download the model checkpoints:
Google Drive Link

Fault Line Detection in Compressive Domain

Download the model checkpoints:
Google Drive Link

---

📜 Abstract

This research introduces a novel compressive domain analytics framework leveraging deep learning to enable efficient seismic image analysis. The framework optimizes for:

• Fault line detection
• Salt trace identification

By utilizing low bits per pixel, the model ensures effective transmission in bandwidth-constrained scenarios, achieving high PSNR and SSIM metrics.

---

🛠️ Key Features

• High Compression Efficiency: Improved seismic data transmission.
• Accurate Fault Detection: Enhanced performance in geophysical analysis.
• Low Resource Requirement: Optimized for compressive domains.

---

---

📬 Contact

For questions or collaboration, please reach out at siddharathnarayan@gmail.com or connect on LinkedIn.