Single Image 3D Object Detection and Pose Estimation for Grasping,
M. Zhu, K. Derpanis, Y. Yang, S. Brahmbhatt, M. Zhang, C. Phillips, M. Lecce and K. Daniilidis, International Conference on Robotics and Automation (ICRA), 2014. [pdf]

In this paper, we address the problem of a robot grasping 3D objects of known 3D shape from their projections in single images of cluttered scenes. We present a novel approach for detecting objects and estimating their 3D pose in single images of cluttered scenes.

From Actemes to Action: A Strongly-supervised Representation for Detailed Action Understanding, W. Zhang, M. Zhu and K. Derpanis, International Conference on Computer Vision (ICCV), 2013. [pdf]

Human action classification (“What action is present in the video?”) and detection (“Where and when is a particular action performed in the video?”) are key tasks for understanding imagery. This paper presents a novel approach for analyzing human actions in non-scripted, unconstrained video settings based on volumetric, x-y-t, patch classifiers, termed actemes.

Monocular Visual Odometry and Dense 3D Reconstruction for On-Road Vehicles,
M. Zhu, S. Ramalingam, Y. Taguchi and T. Garass, European Conference on Computer Vision, workshop on Computer Vision in Vichecle Technology (ECCV), 2012. [pdf]

More and more on-road vehicles are equipped with cameras each day. Accurate ego-motion estimation of a vehicle from video sequences is a challenging and important problem in robotics and computer vision. This paper presents a novel method for estimating the relative motion of a vehicle from a sequence of images obtained using a single vehicle-mounted camera.

Reading texts from natural scene is an easy task for human being but not for a robot. However our PR2 robot, Graspy, became quite smart in this case. In this project, Graspy is programed to be able to detect and recognize the text from an indoor environment. What's more, Graspy is so excited to have this ability that he reads out every single word he sees!

literate_pr2 is now an open source ROS package: http://www.ros.org/wiki/literate_pr2. To learn more about ROS and PR2, take a look at Willow Garage's website