VLSI For Feature Detection and Tracking
Christophe Basset, Bedabrata Pain (JPL), Pietro Perona

We are developing an integrated visual tracking system. The goal of this collaborative work with the Jet Propulsion Laboratory is a single chip serving as a camera (1024x1024 pixels imager array) able to find and track a small (7x7 pixels) target whose image has previously been provided by the user. (full report)

Configurable Architectures, Systems and Tools for Real-Time Low-level Vision
Arrigo Benedetti, Pietro Perona

The long-term goal of this project is to build an infrastructure for the design and implementation of real-time computer vision systems. Since vision algorithms are compute-bound we have chosen the technology of Field Programmable Gate Array (FPGAs), that allow to exploit the instruction level parallelism inherent to the first stages of vision tasks. The first problem that we have considered is the real-time computation of the optical flow measured from the sequence of images captured by a video camera. We have designed, built and demonstrated a system able to select in real-time 2-D visual features on a commercially available reconfigurable platform. During this process we have learned that the system level architectures of off-the-shelf reconfigurable computers are not optimized for low level vision tasks, therefore, we have designed a novel architecture dedicated to real-time processing of video streams. A system based on this architecture has been built and is currently being tested. More recently, we have studied the problem of bit-width computation for the optimization of the data paths found in digital video signal processors. (full report)

Human Action Classification
Xiaolin Feng, Pietro Perona

We study and classify the human actions in this project. We first construct a large dataset of movelets which are defined as body configuration and motion. Each action is represented as the temporal link through the movelets and this temporal link is modeled by Hidden Markov Model. For a given test sequence, the likelihood that it fits the actions we learnt are estimated. The sequence is classified to the action with the maximum likelihood. The algorithm is tested on both periodic (walking, jogging etc.) and nonperiodic (reaching) human actions.
(full report)

Classification of Road Vehicles
Robert Fergus, Bradley Phillips, Paul Updike

We have worked to apply the probabilistic recognition techniques developed in our lab to the classification of road vehicles in busy traffic scenes. We have demonstrated that the model can successfully determine the presence or absence of cars in a given road scene using a detection algorithm that is translation and scale invariant, and can deal with cluttered scenes and occlusions. (full report)

Rapid Natural Scene Categorization without Attention
Fei Fei Li, Rufin VanRullen, Christof Koch, Pietro Perona

Visual attention plays an important role as we walk around the world and recognizes different objects. So what happens when attention is taken away? Are we still able to recognize scenes or objects? Our study finds that certain high level tasks, such as natural scene categorization, can still be performed with little or no attention. (full report)

Shadow Carving
Silvio Savarese, Holly Rushmeier, Fausto Bernardini, Pietro Perona

The shape of an object may be estimated by observing the shadows on its surface. Assuming that a conservative estimate of the object shape is available, our method analyzes images of the object illuminated with known point light sources and taken from known camera locations. The surface estimate is adjusted using the shadow regions to produce a refinement that is still a conservative estimate. A proof of correctness is provided. The method has been tested and validated with experimental results. (full report)

Detection of Human Motion in a Cluttered Scene
Yang Song, Luis Goncalves, and Pietro Perona

Humans are the most important component of a machine's environment. We develop an algorithm which can generate models of human motion automatically from unlabeled real image sequences. Experiments show that the resulting models can successfully detect and label humans from image sequences with clutter and occlusion. (full report)

Primitives for Human Motion: A Dynamical Approach
D. Del Vecchio, R.M. Murray, P. Perona

Using tools from d dynamical systems theory and systems identification theory we develop the study of primitives for human motion which we refer to as movemes. We introduce basic definitions of dynamical independence of LTI systems and segmentability of signals and we develop classification and segmentation algorithms for two dimensional motions. We test our ideas on data sampled from four human subjects who were engaged in a simple real­life activity including two movemes. Our experiments show that we are able to distinguish between the two movemes and recognize them even when they take place in an activity containing more than one moveme.
(full Report)