Hamid Izadinia, PhD

Follow @izadinia

Google citations

PhD Thesis

contact

izadinia AT cs.washington.edu

Computer Science & Engr.
Paul G Allen Center
185 Stevens Way
Seattle, WA 98195-2350

I received PhD degree in Computer Science from University of Washington. I got advice from my PhD advisor Prof. Steve Seitz and my PhD committee Alexei Efros, Byron Boots and Brian Curless . I was also part of the Graphics-Vision Lab (GRAIL) at UW.

My research interest include 3D Deep Learning, Machine Learning, Deep Reinforcement Learning and 3D Scene Understanding.

Work Experience

	Facebook Reality Labs (FRL) Student Researcher Redmond, WA Fall 2017-Winter 2019		Research Intern, Oculus Research San Francisco, CA Summer 2017
	Research Intern, Flickr San Francisco, CA Summer 2016		Research Intern, Adobe Research (Creative Technologies Lab) San Francisco, CA Summer 2014
	Research Intern, Disney Research Pittsburgh, PA Summer 2013		Visiting scholar, Robotics Institute Carnegie Mellon University Pittsburgh, PA Summer 2012

selected publications (All publications)

	Nonprehensile Riemannian Motion Predictive Control Hamid Izadinia, Byron Boots, Steven M. Seitz International Symposium on Experimental Robotics (ISER 2020). Keywords: Riemannian Motion Policy (RMP), Nonprehensile manipulation, Model Predictive Control (MPC), Closed-Loop Control, Real-to-Sim reward analysis, RMPflow, real RC-car robot, fully automatic object-level scene recomposition, point cloud input, YCB dataset. pdf · arXiv · project page
	Scene Recomposition by Learning-based ICP Hamid Izadinia, Steven M. Seitz CVPR 2020. A novel learning-based ICP and fully automatic Scene Recomposition that utilizes thousands of 3D CAD models to align 3D CAD to depth scan by Deep RL. Keywords: 3D scene recomposition, 3D scene reconstruction, Deep reinforcement learning (Deep RL), Learning-based ICP (LICP), 3D geometry learning, 3D CAD models, 3D shapes, Room layout estimation, 3D geometry deep network, Iterative Closest Point, 3D geometry network, noisy scan. pdf · arXiv · project page
	VISER: Visual Self-Regularization Hamid Izadinia, Pierre Garrigues CVPR 2020 - VL3. A novel method to harness real adversarial examples from unlabeled images as a source of regularization data for learning robust visual representation. Keywords: Visual self regularizers, approximate nearest neighbor, real adversarial example, adversarial regularization, regularization for training ConvNets, object-in-context retrieval and visual localization, semi-supervised and weakly-supervised deep learning, fully convolutional deep neural network, t-SNE embedding map, MS COCO and Visual Genome and YFCC100M dataset. pdf · video · project page
	IM2CAD Hamid Izadinia, Qi Shan, Steven M. Seitz CVPR 2017. (spotlight) Keywords: 3D scene reconstruction, convolutional neural network, reconstruction-recognition, scene optimization via render-and-match, single view geometry, room layout estimation, 3D CAD models. pdf · arXiv · project page
	Tag Prediction at Flickr: a View from the Darkroom Kofi Boakye, Sachin Farfade, Hamid Izadinia, Yannis Kalantidis, Pierre Garrigues NeurIPS 2016 - Large Scale Computer Vision systems Workshop. (Best paper award) Keywords: photo tagging, interpretable representation, deep convolutional neural networks, developing large-scale photo tagging system, lightweight and high-performance models, training noisy data. pdf
	Segment-Phrase Table for Semantic Segmentation, Visual Entailment and Paraphrasing H. Izadinia, F. Sadeghi, S.K. Divvala, Y. Choi, A. Farhadi ICCV 2015. (oral) Keywords: visual entailment, visual paraphrasing, semantic segmentation, large-scale recognition. pdf
	Deep Classifiers from Image Tags in the Wild H. Izadinia, B. Russell, A. Farhadi, M. Hoffman, A. Hertzmann Multimedia COMMONS, ACM Multimedia, 2015. Keywords: deep learning, convolutional neural network, large-scale recognition, photo tagging, image classification, image retrieval, robust classification, deep tagging, YFCC100M (as noisy labeled dataset). pdf · Deep-Tagging demo
	Incorporating Scene Context and Object Layout into Appearance Modeling H. Izadinia, F. Sadeghi, A. Farhadi CVPR, 2014. Keywords: scene understanding, large-scale recognition, object layout inference, structured SVM, black-box test, scene context. pdf
	Decoding the Text Encoding F. Sadeghi, H. Izadinia preprint arXiv:1412.6079, December 2014. Keywords: diagram understanding, visualization redesign, text visualization, computer vision, word cloud. pdf
	Recognizing Complex Events using Large Margin Joint Low-level Event Model H. Izadinia, M. Shah ECCV, 2012. Keywords: complex event recognition, action recognition, structured learning, latent SVM. pdf
	(MP)2T: Multiple People Multiple Parts Tracker H. Izadinia, I. Saleemi, W. Li, M. Shah ECCV, 2012. Keywords: multi-target, tracking, pedestrians, humans, body part tracking, network flow optimization, k-shortest paths. pdf
	Multimodal Analysis for Identification and Segmentation of Moving-Sounding Objects H. Izadinia, I. Saleemi, M. Shah IEEE TRANSACTIONS ON MULTIMEDIA, 2013. Keywords: audio-visual analysis, canonical correlation analysis, video segmentation, audio-visual synchronization. pdf
	Multi-Pose Multi-Target Tracking for Human Activity Understanding H. Izadinia, V. Ramakrishna, K. Kitani, D. Huber WACV 2013. Keywords: human pose estimation, multi-target tracking, action recognition. pdf

Teaching GTA

Introduction to Artificial Intelligence CSE 473 (Fall 2018) CSE 415 (Winter 2019)	Reinforcement Learning (RL), Q-Learning, Value Iteration, Markov Decision Process (MDP), CSP, Search, HMM, Bayesian Net, Perceptron.

Misc

Find the slides of our computer vision seminar class CSE590v (Autumn 2014)