Darnell Moore

Exploiting human actions and object context for recognition tasks

Our goal is to exploit human motion and object context to perform action recognition and object classification. Towards this end, we introduce a framework for recognizing actions and objects by measuring image-, object- and action-based information from video. Hidden Markov models are combined with object context to classify hand actions, which are aggregated by a Bayesian classifier to summarize activities. We also use Bayesian methods to differentiate the class of unknown objects by evaluating detected actions along with low-level, extracted object features. Our approach is appropriate for locating and classifying objects under a variety of conditions including full occlusion. We show experiments where both familiar and previously unseen objects are recognized using action and context information.

Implementing phicons: combining computer vision with infrared technology for interactive physical icons

This paper describes a novel physical icon [3] (“phicon”) based system that can be programmed to issue a range of commands about what the user wishes to do with handdrawn whiteboard content. Through the phicons UI, a command to process whiteboard context is issued using infrared signaling in combination with image processing and a ceiling-mounted camera system. We leverage camera systems that are already used for capturing whiteboard content [4] by further augmenting these systems to detect the presence and location of IR beacons within an image. An HDLC-based protocol and a built-in IR transmitter are used to send these signals.

An optimized vision library approach for embedded systems

There is an ever-growing pressure to accelerate computer vision applications on embedded processors for wide-ranging equipment including mobile phones, network cameras, and automotive safety systems. Towards this goal, we propose a software library approach that eases common computational bottlenecks by optimizing over 60 low- and mid-level vision kernels. Optimized for a digital signal processor that is deployed in many embedded image & video processing systems, the library was designed for typical high-performance and low-power requirements. The algorithms are implemented in fixed-point arithmetic and support block-wise partitioning of video frames so that a direct memory access engine can efficiently move data between on-chip and external memory. We highlight the benefits of this library for a baseline video security application, which segments moving foreground objects from a static background. Benchmarks show a ten-fold acceleration over a bit-exact yet unoptimized C language implementation, creating more computational headroom to embed other vision algorithms.

Tracking 3D position and orientation from 2D sequences using simple geometry

This paper introduces a simple way of tracking the position and orientation of objects from a single camera by exploiting the perspective projection model. Three points on the subject that lie in a plane are identified and their distances from the camera lens are measured. Using photometric methods and simple geometry, the location of these points in 3-space is estimated from their projection on the image plane. The direction of orientation can be accurately detected. Promising results using various sequences support the use of this approach in real-time tracking applications.

JPEG2000 for handheld applications

In this paper, we motivate JPEG2000 at length for digital still cameras by highlighting new usage models. These models show how JPEG2000s new features will make digital photography more convenient and practical for handheld applications than JPEG. Some of the major challenges associated with developing cost-effective, high-performance systems are identified. We propose a new block-based approach for wavelet decomposition that reduces memory requirements, enables processing concurrency and efficiency, and promotes coexistence with JPEG.