

439: Deep Learning for Machine Vision
6 snips Jan 28, 2021
Dablina Bhattacharjee, an AI researcher and deep learning engineer, shares her insights into computer vision. She discusses her fascinating work in medical imaging and earthquake prediction, highlighting the importance of foundational mathematics in data science. Dablina also explores the synergy between computer vision and art, revealing innovative projects that enhance diagnostic processes. Listeners can glean valuable advice on essential skills for aspiring data scientists and tips for effective time management while balancing rigorous research and personal pursuits like drumming.
AI Snips
Chapters
Books
Transcript
Episode notes
Plant Intelligence to Medical Images
- Deblina's master's research involved mathematically modeling plant intelligence, like their Fibonacci growth patterns.
- She later applied this search optimization work to detect white blood cells in medical images with high accuracy.
Earthquake Detection
- At Samsung, Deblina used computer vision to detect earthquakes by observing subtle building movements after P-waves hit.
- This work was inspired by MIT research on visual magnification of imperceptible vibrations.
2D to 3D Landscapes
- Deblina's PhD research focuses on converting 2D landscape images (art, photos, comics) into immersive 3D VR experiences.
- The project aims to bridge art, digital humanities, and computer vision, starting with saliency and depth estimation.