In this paper, we propose an algorithm for identifying regions of interest (ROIs) in video, particularly for the keyframes extracted from a home video. The camera motion is introduced as a new factor that can influence the visual saliency. The global motion parameters are used to generate location-based importance maps. These maps can be combined with other saliency maps calculated using other visual and high-level features. Here, we employed the contrast-based saliency as an important low level factor along with face detection as a high level feature in our approach.

This study represents the first cross-cultural investigation of the third person effect hypothesis, which states that individuals overestimate mass media effect on others (Davidson, 1983). It is predicted that the difference between perceived effects of the media on self vs. other will be greater in an individualistic than collectivistic culture, because in the latter self and other are not as separate and the motivation for self-enhancement is not as salient as in the former. Survey data were collected from 671 South Korean (n = 351) and U. S. (n = 320) college students regarding their perceptions about the effects of beer commercials, liquor advertisements, television news about AIDS, and television news about the effects of smoking. The third person effect of undesirable media content emerged from both American and Korean samples, but the size was consistently greater among Americans compared to Koreans. Likewise, the first person effect was greater among Americans rather than Koreans.

This paper describes methods for forensic characterization of imaging devices. This is important in verifying the trust and authenticity of data and the device that created it. Current forensic identification techniques for imaging devices, such as digital cameras, scanners, and printers, are presented. We discuss our ongoing research in creating a forensic framework for these devices. Also discussed are methods we have developed for securely embedding a separate communications channel in the device output.

The HIV/AIDS infection rate of Ethiopia is one of the world’s highest. Prevention campaigns should systematically incorporate and respond to at-risk population’s existing beliefs, emotions, and perceived barriers in the message design process to effectively promote behavior change. However, guidelines for conducting formative evaluation that are grounded in proven risk communication theory and empirical data analysis techniques are hard to find. This article provides a five-step formative evaluation process that translates theory and research for developing effective messages for behavior change. Guided by the extended parallel process model, the five-step process helps message designers manage public’s fear surrounding issues such as HIV/AIDS. An entertainment education project that used the process to design HIV/AIDS prevention messages for Ethiopian urban youth is reported. Data were collected in five urban regions of Ethiopia and analyzed according to the process to develop key messages for a 26-week radio soap opera.

This study represents an initial attempt to examine the intended and unintended effects of fear appeals among individuals in different stages of change. Toward this end, a pilot study investigated the effects of fear appeals promoting skin cancer preventive behavior among college students. After being exposed to fear appeals, individuals who were in the precontemplation stage indicated a greater likelihood of thinking defensively and fatalistically regarding the facts on health risk than those who had intended to engage in or who had previously engaged in preventive behavior. Concurrently, after being exposed to fear appeals, those who were in the precontemplation stage reported less favorable attitudes toward message recommendations, weaker intentions to engage in recommended behavior, and less performance of preventive behavior than those who had contemplated or had previously engaged in preventive behavior. Implications of these results on future public health campaigns are discussed.

This paper describes a method for forensic characterization of RF devices using two-tone probe signals. When transmitted to an RF device, the two-tone signal is affected by nonlinear circuit components such as amplifiers or diodes. The nonlinear components cause intermodulation distortion to the input signal, which is reradiated by the device. Features of the intermodulation distortion products are used to construct a device fingerprint. The fingerprint is then used to characterize the device so that it can be identified from other RF devices.

Many Wyner-Ziv video coding (WZVC) schemes encode a video sequence into two types of frames, key frames and Wyner-Ziv frames. We have previously presented a Wyner-Ziv video coding scheme that uses backward channel aware motion estimation to encode the key frames, where motion estimation was performed at the decoder and motion information was sent back to the encoder. We refer to these backward predictively coded frames as BP frames. In this paper, we extend our previous work and propose three types of motion estimators. A model is presented to examine the complexity-rate-distortion performance of BP frames for the three motion estimators.

We describe three 2D discrete wavelet transform fixed-point implementations and compare them in terms of quantization error for the Daubechies 9/spl times/7 filter bank. The three implementations are the polyphase form, lifting scheme, and reduced scaling lifting scheme. Experimental results show that the reduced scaling lifting scheme is more robust than the other schemes. Also, the number of cycles the implementations take on a Texas Instruments TMS320C6201 simulator are given as reference.

Interleaving is used before the encoding of source symbols in JPEG to reduce visual artifacts due to lost packets because interleaving distributes the locations of errors. The recovery of lost DCT coefficients in interleaved image compression is investigated in this paper. To restore the lost coefficients, an Maximum a Posteriori (MAP) estimate for the DCT coefficients is proposed. Under the assumption of a Gauss-Markov Random Field (GMRF) model in the pixel domain, the MAP estimate for the lost DCT coefficients is derived.

The majority of screening mammograms are normal. It will be beneficial if a detection system is designed to help radiologists readily identify normal regions of mammograms. In this paper, we will present a binary tree classifier based on the use of global features extracted from different levels of a 2-D Quincunx wavelet decomposition of normal and abnormal regional images. This classifier is then used to classify whether an entire whole-field mammogram is normal. This approach is fundamentally different from other approaches that identify a particular abnormality in that is independent of the particular type of abnormality.

Aims: Social norms interventions for reducing excessive drinking on college campuses have reported mixed results. In an attempt to understand the inconsistencies, this study examined some of the under-examined aspects of the social norms approach.

Design: A cross-sectional survey of undergraduate students at two large mid-western universities in the USA was conducted. Norm proximity (campus vs. friends’ norms), norm types (descriptive vs. injunctive norms), and college drinker types (frequent binge drinkers, occasional binge drinkers, and non-binge drinkers) were differentiated.

Findings: The influences of friends’ norms are stronger than those of campus and those of descriptive norms are stronger than injunctive norms. Friends’ descriptive norms influenced frequent and occasional binge drinkers’ behavior most strongly, whereas the campus descriptive norm and self-efficacy influenced non-binge drinkers’ behavior.

Conclusions: Proximity and types of norms as well as types of college drinkers will need to be distinguished in the future design and evaluation of social norms interventions.

In this paper, we present two rate allocation algorithms for embedded motion compensated video coders. The algorithms are based on the modeling of both the video signal and the coder which allow us to express the coding distortion with a recurrence equation. Our algorithms assign rates to the frames of each group of pictures (GOP) of a video sequence in an optimum way. In the first algorithm, the criterion is to minimize the average (MENAVE) distortion and in the second to achieve constant distortion (CD) in all frames. Numerical simulations show the MENAVE criterion can introduce large variations in quality with no significant gains in average distortion with respect to the CD criterion. We also show how the motion estimation accuracy and the GOP length influence in both strategies.

The perceived as well as actual (in) effectiveness of anti-drug ads has been prominent in public discourse and concern. Using the third person effect hypothesis, this study examined adolescents’ perceptions on the effect of anti-drug ads and the perception’s relationship to attitudes and intentions concerning drug use. The results suggested that adolescents estimated the anti-drug ad effect on the basis of their behavioral experience, the self-anchored expectancy of a pro-social media effect: Those who had used drugs anticipated the effect of anti-drug ads to be smaller than those who had not used drugs. The perceived informative realism of anti-drug ads influenced the perceived message quality, which in turn influenced the perceived effect on self. The perceived effect on self was positively associated with anti-drug attitudes and intentions. The implications on future third person perception research and anti-drug campaign efforts are discussed.

Automatic classification of normal mammograms, which constitute a majority of screening mammograms, is a new approach to computer-aided diagnosis of breast cancer. This approach may be limited, however, by non-separable “crossed” distributions of features that are extracted from digitized mammograms. This work presents a method of mapping such non-separable input features into a new set of separable features that can be utilized, together with ordinary “uncrossed” features, by a support vector machine (SVM) classifier. The results of the proposed scheme show improved performance with 80% sensitivity and 95% specificity.