Juha Röning

Using code mobility to create ubiquitous and active augmented reality in mobile computing

The concept augmented reality is used to describe a system that supplements reality by adding virtual objects into a real-world view. This paper describes a flexible mobile code approach that we can use to implement ubiquitous, active, and mobile augmented reality systems. We will concentrate primarily on solving the problem of how to acquire the data for the virtual objects in a way that will be flexible and expandable enough to be used in truly ubiquitous computer systems. To clarify the concept, we will present an example system that provides virtual user interfaces for various real-world objects. This paper suggests that the mobile code approach offers a relatively simple solution that is flexible, scales well, and does not require the computing equipment attached to the real-world objects to be excessively complicated. We can use the same mobile code approach in applications that provide only simple descriptions of the real-world objects and in complex applications that allow the user to manipulate real-world objects via virtual user interfaces that exist only in augmented reality. Our approach also supports active augmented reality, in which the virtual objects can react to the real-world events.

Application of machine vision to assess involved surface in patients with psoriasis.

An objective method is still lacking in the quantitative assessment of psoriasis severity. The purpose of this study was to examine whether computer image analysis (CIA) by a new colour segmentation method can be used as an objective method of estimating involved surface area in patients with psoriasis. Involved surface area in psoriasis was assessed from colour photographs covering the same areas as the psoriasis area and severity index (PASI) scores of 26 patients by human observers from different medical professional groups and by CIA. The reliability of CIA was tested with reference analysis by colouring manually the involved areas in the images with the aid of projected slides. The surface area estimates by the different methods were compared with each other and the resulting effects on the PASI score were assessed. The agreement between the reference analysis and the image analysis was high, especially on trunk areas, but the cylindrical shape of the limbs resulted in some difficulties in assessing involved skin area. The human eye estimates differed from the image analysis ones in almost one‐third of the cases and mostly when psoriasis affected under 30% of the skin surface area. Both over‐and underestimates emerged, but overestimates were more common. Error estimates had a significant effect on the PASI score. The CIA method seems to be reliable and practicable in estimating the actual surface of psoriasis. The method has the disadvantage of being time‐consuming (photographing, processing of pictures) and technically demanding. Further development of this method should make it faster in the future.

Artificial neural networks for aerobic fitness approximation

A unique method for approximating aerobic fitness from demographic and heart rate variables using an artificial neural network (ANN) approximation is proposed. Conventional oxygen uptake measurement methods are expensive and require special clinical instruments. The present method is based on a structure of two ANNs connected in a serial fashion. The first ANN structure is called a preclassifier. It has inputs of physiological features and fuzzy features identified in the material used. The latter ANN structure is a primary approximator, which has an input of the preclassifier result and certain statistical features calculated from the heart rate recordings.

Comparison of clinical and computerized image analyses in the assessment of skin ageing in smokers and non-smokers.

Tobacco smoke and UV radiation are extrinsic risk factors for accelerated skin ageing. In this study the effects of smoking on wrinkling and ageing were assessed in males living in Northern Finland, where cumulative sun exposure is low. Smoking habits, age and facial wrinkling were estimated from facial photographs of 41 smokers and 48 non-smokers by eight panellists, using a blinded standardized assessment. Wrinkling of 26 smokers and 31 non-smokers was also assessed by computerized image analysis. The panellists identified 68% of the smokers correctly as being smokers and the smokers were estimated as being an average of 2.1 years older than their age by the panellists, whereas the non-smokers were estimated as being an average of 0.7 years younger than their age (p < 0.05). No significant difference in skin wrinkling was found between the groups by either clinical assessment or by computerized image analysis. In conclusion, even in the absence of increased wrinkling, the smokers looked older than their age and a majority of them could be identified as smokers by their facial features alone.

CyPhone—bringing augmented reality to next generation mobile phones

We describe a prototype implementation of a future mobile phone called CyPhone. In addition to voice calls, it has been designed to support context-specific and multi-user multimedia services in an augmented reality manner. Context-awareness has been implemented with GPS-based navigation techniques and a registration algorithm, capable of detecting a predefined 3-D model or a landmark in the environment. A new adaptive transport protocol has been developed to support real-time packet-switched data transfer between concurrent users of mobile augmented reality applications. The prototype itself is based on PC/104+ architecture. As a case example we describe an augmented reality-based personal navigation service.

Registration of nevi in successive skin images for early detection of melanoma

The only cure for malignant melanoma (skin cancer) is early detection. Surgical removal of a newly developed melanoma will result in complete cure. In this study, the first steps towards the development of a skin cancer detection system are reported. One possible way to detect early the occurrence of melanoma is to screen the body of a patient at regular intervals for changes or new lesions. To increase the accuracy of this laborious and painstaking task, a computer vision system could be used. One of the most important problems involved in a vision system of this kind is the need to determine which lesions in successive skin images taken over a given period, represent the same lesions. Repeated registrations of skin images also detect new lesions that do not have a counterpart on the previous image. After registration, the lesions in successive images are compared for alterations in size, shape, colour and so on, to detect changes that are suggestive of melanoma. In this paper, we introduce a new algorithm for the registration of lesions in successive skin images. The baseline algorithm requires two initial matches to register the other lesions in the images. The initial matches are provided by a physician or an algorithm that selects the most likely initial matches. The test suggests that the baseline algorithm determines 99% of the matches correctly, and this performance is largely independent of the number of lesions in the skin images.

Comparison of actual psoriasis surface area and the psoriasis area and severity index by the human eye and machine vision methods in following the treatment of psoriasis.

The lack of a quantitative method for assessing psoriasis severity poses a problem for quality control in dermatology. Quantitative estimation of involved surface area is important, as in the psoriasis area and severity index (PASI), but the reliability of many methods is poor. The purpose of this study was to assess the involved surface area of 15 psoriasis patients before and after different anti-psoriasis treatments using the human eye method and a computer image analysis (CIA) system based on colour segmentation. The human eye assessments were compared with the results of the CIA system and the resulting effects on the PASI score were also compared. The human eye estimates were higher than those obtained by the CIA method and, as a consequence, the values of the PASI by the human eye method were also higher than those by CIA. The human eye estimates differed most in cases where the PASI was under 15. The changes in the PASI by the human eye method before and after treatments differed significantly from those by CIA. The CIA system offers a possibility to quantify actual surface in patients with psoriasis, and will be an alternative for developing quality control when evaluating different treatment efficacies.

Reactive task execution by combining action maps

In this paper we describe a behavior-based control architecture capable of reactive task execution; that is, it reasons out actions based on task constraints and reacts quickly to unexpected events in the environment. Reactive task execution requires versatile behavior coordination methods. We suggest that the required behavior coordination can be achieved by combining action maps. The main advantages of this method are that it enables executing several tasks in parallel and facilitates considering environment dynamics. These properties make this behavior coordination method suitable for reactive task execution.

Texture Features in the Classification of Melanocytic Lesions

The use of different texture features for the classification of melanocytic lesions is studied in an attempt to develop a computerized method for the early detection of melanoma. The computer laboratory at the University of Oulu has a strong tradition in applications of computer vision to visual inspection for industrial quality control, and some of the methods learnt in these applications are being transferred and experimented with in this medical context. This include the utilization of texture distributions for classification purposes.

An obstacle detection system using a light stripe identification based method

A light stripe tracking and identification method is proposed for a structured light based obstacle detection system operating in an outdoor environment. The method makes the structured light based detection system more robust and applicable to use outdoors as aid for navigation. The method differentiates between the structured light produced by a light stripe projector and the light stripe kind patterns caused by ambient illumination. The centre of gravity of the segmented light stripe is tracked by using a Kalman filter. The position information together with the other properties of the stripe segment, including intensity, length and orientation are used to identify the same light stripe segment in adjacent images. By using a pulsed light source it is possible to differentiate between true and false light stripes depending on their time of appearance. During the project, a working obstacle detection system for a partly structured outdoor environment was implemented.