Identifying visual cues to enable searching 3D images
2016 Advanced Light Source User Meeting Workshop, 4-5 Oct. 2016, B15, Rm 300
Organizers: Dula Parkinson, Daniela Ushizima, Center for Advanced Mathematics for Energy Research Applications (CAMERA)
Abstract: Many experiments generate images with a large number of visual cues. These can be interpreted by clever and experienced science domain experts, but not by most automated image analysis pipelines. Changes in the instrument or in sample condition all lead to corresponding visual features in the images. How can we automate identification of patterns to separate the signal from the noise, the discovery from the artifact?


CLICK HERE to sign up for "office hours", to be held in Building 6, Room 2244

Tuesday
14:00Dula Parkinson, LBNLWelcome
14:05Dani Ushizima, LBNLCAMERA, computer vision and scientific images
14:35Chao Yang, LBNLNeuromorphic image processing
15:05Break
15:25Stella Yu, UCBAffinity CNN: Learning Pixel-Centric Pairwise Relations for Image Segmentation
15:55Daniel Pelt, LBNLOptimizing tomographic reconstruction for specific analysis tasks
16:25Stefan van del Walt, UCB"Python, Scikit-image, and new tools for image analysis"
16:55End
Wednesday
10:00Andrew McElrone, UCD
10:30Paulo Monteiro, UCBA multi-scale approach to image cracks, microcracks, and nanoporosity in cementitious material
11:00Natalie Larson, UCSB/ALSTitle: Feature recognition in X-ray tomography data of Ceramic Matrix Composites
11:30Marco Voltolini, LBNL"When thresholding does not work: the separation of objects by their morphology. Case studies with cement, volcanic rocks, and seal rocks".
12:00lunch
13:30"office hours"Participants can schedule time to meet with the speakers to discuss problems and solutions and possible collaborations.