This thesis explores the contactless estimation of people’s vital signs. We designed two camera-based systems and applied object detection algorithms to locate the regions of interest where vital signs are estimated. With the development of Deep Learning, Convolutional Neural Network (CNN) model has many applications in the real world nowadays. We applied the CNN based frameworks to the different types of camera based systems and improve the efficiency of the contactless vital signs estimation. In the field of medical healthcare, contactless monitoring draws a lot attention in the recent years because the wide use of different sensors. However most of the methods are still in the experimental phase and have never been used in real applications. We were interested in monitoring vital signs of patients lying in bed or sitting around the bed at a hospital. This required using sensors that have range of 2 to 5 meters. We developed a system based on the depth camera for detecting people’s chest area and the radar for estimating the respiration signal. We applied a CNN based object detection method to locate the position of the subject lying in the bed covered with blanket. And the respiratory-like signal is estimated from the radar device based on the detected subject’s location. We also create a manually annotated dataset containing 1,320 depth images. In each of the depth image the silhouette of the subject’s upper body is annotated, as well as the class. In addition, a small subset of the depth images also labeled four keypoints for the positioning of people’s chest area. This dataset is built on the data collected from the anonymous patients at the hospital which is substantial. Another problem in the field of human vital signs monitoring is that systems seldom contain the functions of monitoring multiple vital signs at the same time. Though there are few attempting to work on this problem recently, they are still all prototypes and have a lot limitations like shorter operation distance. In this application, we focused on contactless estimating subjects’ temperature, breathing rate and heart rate at different distances with or without wearing the mask. We developed a system based on thermal and RGB camera and also explore the feasibility of CNN based object detection algorithms to detect the vital signs from human faces with specifically defined RoIs based on our thermal camera system. We proposed the methods to estimate respiratory rate and heart rate from the thermal videos and RGB videos. The mean absolute difference (MAE) between the estimated HR using the proposed method and the baseline HR for all subjects at different distances is 4.24 ± 2.47 beats per minute, the MAE between the estimated RR and the reference RR for all subjects at different distances is 1.55 ± 0.78 breaths per minute.

A dissertation submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of requirements for the degree of Master of Science. Degree awarded with distinction on 4 December 1996. Johannesburg 1996. / The problem with existing traffic light systems is that they are static and unable to respond to dynamic changes in traffic flow patterns. This report discusses an approach using distributed AI techniques to allow the control system to respond dynamically to the changes in the flow patterns. A simulated system (Datilis) has been developed, The design of the system is based on the assumption that maximising throughput at individual intersections will result in improved overall performance. The results generated and discussed in this report support this hypothesis.

Classic studies found advantages of pictures over words in memory (picture superiority effect). This paper applied the picture superiority effect to road signs to examine whether people remembered picture road signs better than word road signs. People remembered picture road signs better than word road signs, as evidenced on a recognition test. However, in real driving situations drivers do not need to remember the meaning of one road sign for a long time; rather, they need to continuously update information from sequentially encountered road signs. Therefore, Experiment 2 explored the differences in updating memory between pictures and words. Memory for the most recently viewed road signs was different depending on the form (picture, word) of the previously encountered road sign. Previously encountered picture items impaired memory significantly more than previously encountered words. These findings demonstrate that superior picture memory sometimes can be detrimental, especially when remembering recent information.

No description available.

Thesis (MPhil (Philosophy))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: This study investigated the practice of language use on road signs in Switzerland. The language used in public spaces constitutes our image of a specific place to a remarkable degree. We are constantly exposed to signs, whether road signs, advertisements or notifiers. These signs not only convey information such as “Caution”, “Final sale” or “Speed limit 60 here”, but also highlight aspects of culture and, where written text appear on them, aspects of the language(s) used in that area. Language use within the public space enables us to conceptualize, as well as recognize, the ways in which language use occurs, often (but not always) informing us whether a country is monolingual, bilingual or multilingual. In addition, these signs also indicate how language use is linked to other processes such as globalisation, social change and cultural heritage. The main purpose of this thesis was to establish how language is used on road signs in presentday, officially multilingual Switzerland. In this regard, three Swiss cantons (in which bilingual or multilingual societies are found) were investigated as to the language(s) – or lack thereof – which appear on their road signs, as well as to the order in which these languages appear. Specifically, bilingual or multilingual signs, as well as so-called Bildersprache (“picture language”, in other words on a board or plaque with a picture, symbol or other image on it), were studied in these cantons. Data for this research consists of photographic material (taken during visits to the study areas) as well as interview data. It was found that the social and political identities, as well as the economic interests, of language groups in a particular area affect the design and placement of road signs in that area. The results of this study serve as a framework for future research, and the photo material enables us, to a certain extent, to see how an apparently stable language system (such as the multilingualism found in Switzerland) encounters changes due to internal and external influences, but also how creative language can be. / AFRIKAANSE OPSOMMING: In hierdie studie is die taalgebruik op padtekens in Switserland ondersoek. Die taalgebruik in openbare plekke vorm tot ’n merkwaardige mate die beeld wat ons van ’n spesifieke plek het. Ons word konstant blootgestel aan tekens, hetsy padtekens, advertensies of kennisgewings. Hierdie tekens verskaf nie net inligting soos “Versigtig”, “Finale uitverkoping” of “Spoedgrens 60” nie, maar beklemtoon ook kulturele aspekte en, in daardie gevalle waarin geskrewe teks op tekens verskyn, ook aspekte van taal/tale wat in die betrokke gebied gebesig word. Taalgebruik in die openbare ruimte help ons om die wyses waarop taalgebruik voorkom te konseptualiseer en te herken, en dui gereeld (maar nie altyd nie) vir ons aan of ’n land eentalig, tweetalig of veeltalig is. Verder gee die tekens ook ’n aanduiding van hoe taalgebruik gekoppel is aan prosesse soos globalisasie, sosiale verandering en kulturele erfenis. Die hoof doel van hierdie tesis was om te bepaal hoe taal in die hedendaagse Switzerland op padtekens gebruik word. Drie Switserse kantons (met tweetalige en veeltalige gemeenskappe) is in hierdie verband ondersoek in terme van die taal/tale – of die gebrek daaraan – wat op padtekens verskyn asook die volgorde waarin hierdie tale op die padtekens verskyn. Spesifiek tweetalige of veeltalige tekens asook sogenaamde Bildersprache (“prentetaal”, met ander woorde ’n bord met ’n prent, simbool of ander voorstelling daarop) is in hierdie kantons bestudeer. Die data vir hierdie navorsing bestaan uit fotografiese materiaal (wat verkry is tydens besoeke aan die studie-areas) sowel as inligting wat verkry is deur onderhoudvoering. Daar is bevind dat die sosiale en politiese identiteite, sowel as die ekonomiese belange, van taalgroepe in ’n spesifieke area die ontwerp en plasing van padtekens in daardie area beïnvloed. Die resultate van hierdie studie dien as ’n raamwerk vir toekomstige navorsing en die fotomateriaal stel ons tot ’n sekere mate in staat om vas te stel hoe ’n klaarblyklik stabiele taalsisteem (soos die veeltalige een wat in Switserland aangetref word) veranderinge weens interne en eksterne invloede kan ervaar, maar ook hoe kreatief taal kan wees.

An intensive examination was conducted to test the credibility of current traffic signal change interval policies founded on a kinematic equation developed nearly 30 years ago. The investigation involved the review of relevant literature as well as an extensive collection of data. The literature review and data analysis revealed that current change interval policies rely on the disproven assumption that traffic decelerates at a constant rate. The data analysis also demonstrated that traffic approach speed and deceleration distance affect the manner in which deceleration occurs. Based on the data analysis, an alternative treatment of the kinematic equation is proposed using surrogate deceleration rates. The surrogate rates offer a pragmatic set of input for the kinematic equation. Therefore, rather than yielding a change interval based on an inaccurate assumption, agencies can implement change intervals which are responsive to local traffic.

published_or_final_version / abstract / Conservation / Master / Master of Science in Conservation

published_or_final_version / English Studies / Master / Master of Arts

published_or_final_version / Transport Policy and Planning / Master / Master of Arts in Transport Policy and Planning

This thesis investigates the emergence in human evolution of communication through symbols, or conventional, arbitrary signs. Previous work has argued that symbolic speech was preceded by communication through nonarbitrary signs, but how vocal symbolic communication arose out of this has not been extensively studied. Thus far, past research has emphasized the advantages of vocal symbols and pointed to communicative and evolutionary pressures that would have spurred their development. Based on semiotic principles, I examine emergence in terms of two factors underlying symbols: interpretation and conventionalization. I address the question with a consideration of embodied human experience – that is, accounting for the particular features that characterize human communication. This involves simultaneous expression through vocal and gestural modalities, each of which has distinct semiotic properties and serves distinct functions in language today. I examine research on emerging sign systems together with research on properties of human communication to address the question of symbol emergence in terms of the specific context of human evolution. I argue that, instead of in response to pressures for improved communication, symbolic vocalizations could have emerged through blind cultural processes out of the conditions of multimodal nonarbitrary communication in place prior to modern language. Vocalizations would have been interpreted as arbitrary by virtue of their semiotic profile relative to that of gesture, and arbitrary vocalizations could have become conventionalized via the communicative support of nonarbitrary gestures. This scenario avoids appealing to improbable evolutionary and psychological processes and provides a comprehensive and evolutionarily sound explanation for symbol emergence. I present experiments that test hypotheses stemming from this claim. I show that novel arbitrary vocal forms are interpreted and adopted as symbols even when these are uninformative and gesture is the primary mode of communication. I also present computational models that simulate multi-channel, heterosemiotic communication like that of arbitrary speech and nonarbitrary gesture. These demonstrate that information like that provided by gesture can enable the conventionalization of symbols across a population. The results from experiments and simulations together support the claim that symbolic communication could arise naturally from multimodal nonarbitrary communication, offering an explanation for symbol emergence more consistent with evolutionary principles than existing proposals.