Kristin Dana

The Computer Vision and Robotics Laboratory, conducts innovative research at the intersection of computer vision and robotics. Focused on areas in machine learning, artificial intelligence and computational photography, the lab is developing new methods in the application areas of precision agriculture, remote sensing, and socially cognizant robotics. Project examples include visual navigation, change detection in satellite imagery, drone-based cranberry crop evaluation, pedestrian behavior analysis, novel cameras for BRDF measurement, photographic steganography, and quantitative dermatology. The lab's interdisciplinary approach reflects the lab's commitment to exploring and understanding the multifaceted applications of AI in our world.

Select Research Projects

SOCRATES

Socially Cognizant Robotics For a Technology Enhanced Society
Project Page

Vision on the Bog

Finding Berries
Project Page

Multimodal: Vision and Tactile

Teaching Cameras to Feel
Project Page ACM TechNews and New Scientist Article

Remote Sensing

Material Segmentation of Multi-View Satellite Imagery
Project Page

Urban Semantic 3D Reconstruction From Multiview Satellite Imagery
Project Page

Texture and Material Recognition

Deep TEN: Texture Encoding Network
Project Page

Differential Angular Imaging for Material Recognition
Project Page
Ground Terrain Dataset

Friction from Reflectance: Deep Reflectance Codes for Predicting Physical Surface Properties from One-Shot In-Field Reflectance
Hang Zhang, Kristin Dana, Ko Nishino
European Conference on Computer Vision (ECCV), 2016

Reflectance Hashing for Material Recognition
Hang Zhang, Kristin Dana, Ko Nishino
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015

Quantitative Dermatology

Skin Texture Modeling
Project Page

Software & Datasets

Filename	Description	Size
Deep-Encoding-master.zip	Software toolkit: Torch Impementation of Encoding Layer (Paper 2017). Reproducting the experimental results of Deep-TEN.	42K
Deep-Reflectance-Code-master.zip	Software toolkit: Reproducting the experimental results of DRC (ECCV 2016). Hashing for material recognition and friction estimation. Baseline approaches for image retrieval.	214K
RF2016.zip	The package contains: Reflectance disks of 137 materials (5754 images). The measured coefficient of kinetic friction of each material sample.	1.15G

Publications

See google scholar page.