Comparing segmentation strategies for efficient video passage retrieval

We compare the effect of different text segmentation strategies on speech based passage retrieval of video. Passage retrieval has mainly been studied to improve document retrieval and to enable question answering. In these domains best results were obtained using passages defined by the paragraph structure of the source documents or by using arbitrary overlapping passages. For the retrieval of relevant passages in a video, using speech transcripts, no author defined segmentation is available. We compare retrieval results from 4 different types of segments based on the speech channel of the video: fixed length segments, a sliding window, semantically coherent segments and prosodic segments. We evaluated the methods on the corpus of the MediaEval 2011 Rich Speech Retrieval task. Our main conclusion is that the retrieval results highly depend on the right choice for the segment length. However, results using the segmentation into semantically coherent parts depend much less on the segment length. Especially, the quality of fixed length and sliding window segmentation drops fast when the segment length increases, while quality of the semantically coherent segments is much more stable. Thus, if coherent segments are defined, longer segments can be used and consequently less segments have to be considered at retrieval time.

Designing personalized video-based crossmedia informal learning environments beyond iTV

Tese de doutoramento, Informática (Engenharia Informática), Universidade de Lisboa, Faculdade de Ciências, 2014

Beach morphodynamics at Nazaré coast using video monitoring

Tese de doutoramento, Geologia (Geodinâmica Externa), Universidade de Lisboa, Faculdade de Ciências, 2014

Short films about flying

Online artwork which streams web-cam images live from the Internet and re-mixes them into disjointed narrative sequences, thereby producing cinema as a 'found object' made entirely of live material streamed from the internet. ‘Short Films about Flying’ is an online film which explores how a cinematic work can be generated using live material from the internet. The work is driven by software that takes surveillance video from a live camera feed at Logan Airport, Boston, and combines this with randomly grabbed audio from the web and texts taken from websites, chat rooms, message boards etc. This results in an endless open edition of unique cinematic works in real-time. By combining the language of cinema with global real-time data technologies, this work is one of the first new media artworks to re-imagine the internet in a different sensory form as a cinematic space. ‘Short Films about Flying’ was developed over the course of a year in collaboration with Jon Thomson (Slade) to explore how the concept of the found object can be re-conceptualised as the found data stream. It has informed other research by Craighead and Thomson, such as the web project http://www.templatecinema.com, and began an examination into relationships between montage and live virtual data –an early example of which would be ‘Flat Earth’, an animated work developed for Channel 4 in 2007, with the production company Animate. This piece has been cited in discussions on new media art, as a significant example of artworks using a database as their determining structure. It was acquired for the Arts Council Collection and has continuously toured significant international venues over the last 4 years. Citations include:’ Time and Technology’ by Charlie Gere (2006); 'The Wrong Categories' by Kris Cohen (2006); 'Networked Art - Practices and Positions' edited by Tom Corby (Routledge 2005) and Grayson Perry in The Times (9.8.06).

Automatic Video Analysis for Fisheries Survey Systems

Thesis (Ph.D.)--University of Washington, 2015

Evaluation and Potential Cost-Effectiveness of Active Surveillance Pharmacovigilance for First-Line HAART in Namibia

Thesis (Ph.D.)--University of Washington, 2016-02

(un)childhood: performing the voices and times of childhood through relational video-making

This practice-based PhD is comprised of two interrelated elements: (i) ‘(un)childhood’, a 53’ video-essay shown on two screens; and (ii) a 58286 word written thesis. The project, which is contextualised within the tradition of artists working with their own children on time-based art projects, explores a new approach to timebased artistic work about childhood. While Stan Brakhage (1933-2003), Ernie Gher (1943-), Erik Bullot (1963-) and Mary Kelly (1941-) all documented, photographed and filmed their children over a period of years to produce art projects (experimental films and a time-based installation), these projects were implicitly underpinned by a construction of childhood in which children, shown as they grow, represent the abstract primitive subject. The current project challenges the convention of representing children entirely from the adult’s point of view, as aesthetic objects without a voice, as well as through the artist’s chronological approach to time. Instead, this project focuses on the relational joining of the child’s and adult’s points of view. The artist worked on a video project with her own son over a four-and-a-half year period (between the ages of 5 and 10) through which she developed her ‘relational video-making’ methodology. The video-essay (un)childhood performs the relational voices of childhood as resulting from the verbal interactions of both children and adults. The non-chronological nature of(un)childhood offers an alternative to the linear-temporal approach to the representation of childhood. Through montage and a number of literal allusions to time in its dialogue, (un)childhood performs the relational times of childhood by combining children’s lives in the present with the temporal dimensions that have traditionally constructed childhood: past, future and timeless.

Special Section Guest Editorial: Image/Video Quality and System Performance

Rapid developments in display technologies, digital printing, imaging sensors, image processing and image transmission are providing new possibilities for creating and conveying visual content. In an age in which images and video are ubiquitous and where mobile, satellite, and three-dimensional (3-D) imaging have become ordinary experiences, quantification of the performance of modern imaging systems requires appropriate approaches. At the end of the imaging chain, a human observer must decide whether images and video are of a satisfactory visual quality. Hence the measurement and modeling of perceived image quality is of crucial importance, not only in visual arts and commercial applications but also in scientific and entertainment environments. Advances in our understanding of the human visual system offer new possibilities for creating visually superior imaging systems and promise more accurate modeling of image quality. As a result, there is a profusion of new research on imaging performance and perceived quality.

Argonauta Video Database: General simulated scenes. Scene 1 (Right camera)

General simulated scenes These scenes followed a pre-defined script (see the Thesis for details), with common movements corresponding to general experiments. People go to or stand still in front of "J9", and/or go to the side of Argonauta reactor and come back again. The first type of movement is common during Irradiation experiments, where a material sample is put within the "J9" channel; and also during neutrongraphy or gammagraphy experiments, where a sample is placed in front of "J9". Here, the detailed movements of putting samples on these places were not reproduced in details, but only the whole bodies' movements were simulated (as crouching or being still in front of "J9"). The second type of movement may occur when operators go to the side of Argonauta to verify some operational condition. - Scene 1 (Obs.: Scene 1 of the "General simulated scenes" class): Comprises one of the scenes with two persons. Both of them use clothes of light colors. Both persons remain still in front of "J9"; one goes to the computer and then come back, and both go out. Video file labels: "20140326145315_IPCAM": recorded by the right camera,

Argonauta Video Database: General Simulated. Scene 1 (Left Camera)

General simulated scenes These scenes followed a pre-defined script (see the Thesis for details), with common movements corresponding to general experiments. People go to or stand still in front of "J9", and/or go to the side of Argonauta reactor and come back again. The first type of movement is common during Irradiation experiments, where a material sample is put within the "J9" channel; and also during neutrongraphy or gammagraphy experiments, where a sample is placed in front of "J9". Here, the detailed movements of putting samples on these places were not reproduced in details, but only the whole bodies' movements were simulated (as crouching or being still in front of "J9"). The second type of movement may occur when operators go to the side of Argonauta to verify some operational condition. - Scene 1 (Obs.: Scene 1 of the "General simulated scenes" class): Comprises one of the scenes with two persons. Both of them use clothes of light colors. Both persons remain still in front of "J9"; one goes to the computer and then come back, and both go out. Video file labels: "20140326145316_IPCAM": recorded by the left camera.

Argonauta Video Database: Spectrography experiment. Scene 1 (Right câmera)

Scenes for Spectrography experiment Scenes were recorded following the tasks involved in spectrography experiments, which are carried out in front of "J9" output radiadion channel, the latter in open condition. These tasks may be executed by one or two persons. One person can do the tasks, but requiring him to crouch in front of "J9" to adjust the angular position the experimental appartus (a crystal to bend the neutron radiation to the spectograph), and then to get up to verify data in a computer aside; these movements are repeated until achieving the right operational conditions. Two people may aid one another in such a way one remais crouched while the other remains still in front of the computer. They may also interchange tasks so as to divide received doses. Up to now, there are available two scenes with one person and one scene with two persons. These scenes are described in the sequel: - Scene 1: Comprises one of the scenes with one person performing spectography experiment. Video file labels: "20140327181335_IPCAM": recorded by the right camera

Argonauta Video Database: General simulated. Scene 2 (Right camera)

General simulated scenes These scenes followed a pre-defined script (see the Thesis for details), with common movements corresponding to general experiments. People go to or stand still in front of "J9", and/or go to the side of Argonauta reactor and come back again. The first type of movement is common during Irradiation experiments, where a material sample is put within the "J9" channel; and also during neutrongraphy or gammagraphy experiments, where a sample is placed in front of "J9". Here, the detailed movements of putting samples on these places were not reproduced in details, but only the whole bodies' movements were simulated (as crouching or being still in front of "J9"). The second type of movement may occur when operators go to the side of Argonauta to verify some operational condition. - Scene 2: Comprises one of the scenes with two persons. Both of them use clothes of dark colors. Both persons go to the side of Argonauta reactor and then come back and go out. Video file labels: "20140326154754_IPCAM": recorded by the right camera.

Argonauta Video Database: Spectography experiment. Scene 1 (Left camera)

Scenes for Spectrography experiment Scenes were recorded following the tasks involved in spectrography experiments, which are carried out in front of "J9" output radiadion channel, the latter in open condition. These tasks may be executed by one or two persons. One person can do the tasks, but requiring him to crouch in front of "J9" to adjust the angular position the experimental appartus (a crystal to bend the neutron radiation to the spectograph), and then to get up to verify data in a computer aside; these movements are repeated until achieving the right operational conditions. Two people may aid one another in such a way one remais crouched while the other remains still in front of the computer. They may also interchange tasks so as to divide received doses. Up to now, there are available two scenes with one person and one scene with two persons. These scenes are described in the sequel: - Scene 1: Comprises one of the scenes with one person performing spectography experiment. Video file labels:"20140327181336_IPCAM": recorded by the left camera.

Argonauta Video Database: Spectography experiment. Scene 2 (Right camera)

Scenes for Spectrography experiment Scenes were recorded following the tasks involved in spectrography experiments, which are carried out in front of "J9" output radiadion channel, the latter in open condition. These tasks may be executed by one or two persons. One person can do the tasks, but requiring him to crouch in front of "J9" to adjust the angular position the experimental appartus (a crystal to bend the neutron radiation to the spectograph), and then to get up to verify data in a computer aside; these movements are repeated until achieving the right operational conditions. Two people may aid one another in such a way one remais crouched while the other remains still in front of the computer. They may also interchange tasks so as to divide received doses. Up to now, there are available two scenes with one person and one scene with two persons. These scenes are described in the sequel: - Scene 2: Another take similat to Scene 1. Video file labels: "20140327180749_IPCAM": recorded by the right camera.