Pamela C. Cosman

Human Body Model Acquisition and Tracking Using Voxel Data

We present an integrated system for automatic acquisition of the human body model and motion tracking using input from multiple synchronized video streams. The video frames are segmented and the 3D voxel reconstructions of the human body shape in each frame are computed from the foreground silhouettes. These reconstructions are then used as input to the model acquisition and tracking algorithms.The human body model consists of ellipsoids and cylinders and is described using the twists framework resulting in a non-redundant set of model parameters. Model acquisition starts with a simple body part localization procedure based on template fitting and growing, which uses prior knowledge of average body part shapes and dimensions. The initial model is then refined using a Bayesian network that imposes human body proportions onto the body part size estimates. The tracker is an extended Kalman filter that estimates model parameters based on the measurements made on the labeled voxel data. A voxel labeling procedure that handles large frame-to-frame displacements was designed resulting in very robust tracking performance.Extensive evaluation shows that the system performs very reliably on sequences that include different types of motion such as walking, sitting, dancing, running and jumping and people of very different body sizes, from a nine year old girl to a tall adult male.

Evaluating quality of compressed medical images: SNR, subjective rating, and diagnostic accuracy

Compressing a digital image can facilitate its transmission, storage, and processing. As radiology departments become increasingly digital, the quantities of their imaging data are forcing consideration of compression in picture archiving and communication systems (PACS) and evolving teleradiology systems. Significant compression is achievable only by lossy algorithms, which do not permit the exact recovery of the original image. This loss of information renders compression and other image processing algorithms controversial because of the potential loss of quality and consequent problems regarding liability, but the technology must be considered because the alternative is delay, damage, and loss in the communication and recall of the images. How does one decide if an image is good enough for a specific application, such as diagnosis, recall, archival, or educational use? The authors describe three approaches to the measurement of medical image quality: signal-to-noise ratio (SNR), subjective rating, and diagnostic accuracy. They compare and contrast these measures in a particular application, consider in some depth recently developed methods for determining diagnostic accuracy of lossy compressed medical images and examine how good the easily obtainable distortion measures like SNR are at predicting the more expensive subjective and diagnostic ratings. The examples are of medical images compressed using predictive pruned tree-structured vector quantization, but the methods can be used for any digital image processing that produces images different from the original for evaluation. >

Moving shadow and object detection in traffic scenes

We present an algorithm for segmentation of traffic scenes that distinguishes moving objects from their moving cast shadows. A fading memory estimator calculates mean and variance of all three color components for each background pixel. Given the statistics for a background pixel, simple rules for calculating its statistics when covered by a shadow are used. Then, MAP classification decisions are made for each pixel. In addition to the color features, we examine the use of neighborhood information to produce smoother classification. We also propose the use of temporal information by modifying class a priori probabilities based on predictions from the previous frame.

Vector quantization of image subbands: a survey

Subband and wavelet decompositions are powerful tools in image coding because of their decorrelating effects on image pixels, the concentration of energy in a few coefficients, their multirate/multiresolution framework, and their frequency splitting, which allows for efficient coding matched to the statistics of each frequency band and to the characteristics of the human visual system. Vector quantization (VQ) provides a means of converting the decomposed signal into bits in a manner that takes advantage of remaining inter and intraband correlation as well as of the more flexible partitions of higher dimensional vector spaces. Since 1988, a growing body of research has examined the use of VQ for subband/wavelet transform coefficients. We present a survey of these methods.

Modeling packet-loss visibility in MPEG-2 video

We consider the problem of predicting packet loss visibility in MPEG-2 video. We use two modeling approaches: CART and GLM. The former classifies each packet loss as visible or not; the latter predicts the probability that a packet loss is visible. For each modeling approach, we develop three methods, which differ in the amount of information available to them. A reduced reference method has access to limited information based on the video at the encoders side and has access to the video at the decoders side. A no-reference pixel-based method has access to the video at the decoders side but lacks access to information at the encoders side. A no-reference bitstream-based method does not have access to the decoded video either; it has access only to the compressed video bitstream, potentially affected by packet losses. We design our models using the results of a subjective test based on 1080 packet losses in 72 minutes of video.

Automatic tracking, feature extraction and classification of C. elegans phenotypes

This paper presents a method for automatic tracking of the head, tail, and entire body movement of the nematode Caenorhabditis elegans (C. elegans) using computer vision and digital image analysis techniques. The characteristics of the worms movement, posture and texture information were extracted from a 5-min image sequence. A Random Forests classifier was then used to identify the worm type, and the features that best describe the data. A total of 1597 individual worm video sequences, representing wild type and 15 different mutant types, were analyzed. The average correct classification ratio, measured by out-of-bag (OOB) error rate, was 90.9%. The features that have most discrimination ability were also studied. The algorithm developed will be an essential part of a completely automated C. elegans tracking and identification system.

Statistical channel knowledge-based optimum power allocation for relaying protocols in the high SNR regime

We are concerned with transmit power optimization in a wireless relay network with various cooperation protocols. With statistical channel knowledge (in the form of knowledge of the fading distribution and the path loss information across all the nodes) at the transmitters and perfect channel state information at the receivers, we derive the optimal power allocation that minimizes high signal-to-noise ratio (SNR) approximations of the outage probability of the mutual information (MI) with amplify-and-forward (AF), decode-and-forward (DF) and distributed space-time coded (DSTC) relaying protocols operating over Rayleigh fading channels. We demonstrate that the high SNR approximation-based outage probability expressions are convex functions of the transmit power vector, and the nature of the optimal power allocation depends on whether or not a direct link between the source and the destination exists. Interestingly, for AF and DF protocols, this allocation depends only on the ratio of mean channel power gains (i.e., the ratio of the source-relay gain to the relay-destination gain), whereas with a DSTC protocol this allocation also depends on the transmission rate when a direct link exists. In addition to the immediate benefits of improved outage behavior, our results show that optimal power allocation brings impressive coding gains over equal power allocation. Furthermore, our analysis reveals that the coding gain gap between the AF and DF protocols can also be reduced by the optimal power allocation.

Combined forward error control and packetized zerotree wavelet encoding for transmission of images over varying channels

One method of transmitting wavelet based zerotree encoded images over noisy channels is to add channel coding without altering the source coder. A second method is to reorder the embedded zerotree bitstream into packets containing a small set of wavelet coefficient trees. We consider a hybrid mixture of these two approaches and demonstrate situations in which the hybrid image coder can outperform either of the two building block methods, namely on channels that can suffer packet losses as well as statistically varying bit errors.

Using machine vision to analyze and classify Caenorhabditis elegans behavioral phenotypes quantitatively.

Mutants with abnormal patterns of locomotion, also known as uncoordinated (Unc) mutants, have facilitated the genetic dissection of many important aspects of nervous system function and development in the nematode Caenorhabditis elegans. Although a large number of distinct classes of Unc mutants can be distinguished by an experienced observer, precise quantitative definitions of these classes have not been available. Here we describe a new approach for using automatically-acquired image data to quantify the locomotion patterns of wild-type and mutant worms. We designed an automated tracking and imaging system capable of following an individual animal for long time periods and saving a time-coded series of digital images representing its motion and body posture over the course of the recording. We have also devised methods for measuring specific features from these image data that can be used by the classification and regression tree classification algorithm to reliably identify the behavioral patterns of specific mutant types. Ultimately, these tools should make it possible to evaluate with quantitative precision the behavioral phenotypes of novel mutants, gene knockout lines, or pharmacological treatments.

Using vector quantization for image processing

A review is presented of vector quantization, the mapping of pixel intensity vectors into binary vectors indexing a limited number of possible reproductions, which is a popular image compression algorithm. Compression has traditionally been done with little regard for image processing operations that may precede or follow the compression step. Recent work has used vector quantization both to simplify image processing tasks, such as enhancement classification, halftoning, and edge detection, and to reduce the computational complexity by performing the tasks simultaneously with the compression. The fundamental ideas of vector quantization are explained, and vector quantization algorithms that perform image processing are surveyed. >