Fig 1 - uploaded by Fernando Moita
Content may be subject to copyright.
Source publication
This paper presents EyeSEC, a project being conducted in DEE/ISEC in the area of Assistive Technology, which aims to create alternative technologies to compensate for functional limitations and facilitate the independence of elderly and impaired users. The main goal is to merge the data from eye-tracking, provided by vision sensors embedded in eyeg...
Contexts in source publication
Context 1
... in head mounted mode. Since the webcam is placed near the eye it is not required high resolution, but the device dimension should be small in order to minimize the impact it will have on the user's field of vision and weight of the prototype. As a support structure for the webcam and inertial sensors it is proposed to use a frame of eyeglasses (Fig. 1). There are several software programs to detect eye movements, but they are too expensive [3] or complex [4] for our proposed application. For that reason, new processing software was developed under the scope of a BSc project running at DEE/ISEC. ...
Context 2
... Kalman algorithm used to filter the extracted position data (x c , y c ), and presented in the last section, provides the estimation for the new ROI location and R values (Fig. ...
Context 3
... several small objects removed by erosion (Fig. 9b). The circumference of the pupil will not close completely preventing the implementation of geometric algorithms previously mentioned. Therefore, it can be classified by the HT method, obtaining a 50% score. Even in the presence of noisy and open circunferences, the Hough transform is effective (Fig. 10). As the area of image processing is restricted by the predicted ROI, the probability of finding another circumference, even without geometric proofs, is very small. Moreover, these results are applied to a Kalman filter that, according to the model will avoid large trajectory deviations. Table 1 shows the eye detection rate of errors ...
Context 4
... involves a cycle of prediction, observation, and updating (there is often an additional data validation step prior to updating). The flowchart in Figure 11 summarizes the Kalman filter process. In the last step is calculated the estimate that will be used to determine the ROI position in the next frame. ...
Context 5
... validate the eye tracking system, a trajectory was planned on a computer screen. As the image moved along this path at a constant speed, the user must follow it with his eyes wearing the eye-tracker. The first tracking experiment was carried out using a third party open source software. Fig. 12 presents the filter results for a loop controlled by eye-tracking movements on a 1280,800 resolution screen. As can be observed, the raw data is very irregular in time, highly noisy and frequently jump off the ...
Similar publications
The increasing number of older individuals, especially those in the oldest-old age group, has resulted in a significant increase in the number of people with disabilities. Assistive technology offers the potential of increasing independence and quality of life for older individuals with disabilities, as well as reducing health-related costs. To inc...
The purpose of this study was to survey the homes and assistive technology of home-bound elderly people in Namprae sub-district, Hang Dong district, Chiang Mai, Thailand. The home evaluation included features inside and outside the home, and assistive technology was based on the International Classification of Functioning (ICF) concept. Methods, ob...
Citations
... Therefore, by using the previous frame's information (COG and posture of the target) to determine the ROI for the current frame, we can reduce surrounding image noise and processing cost simultaneously. Although similar techniques have been used in other works with high-speed vision systems [22][23][24], these systems did not have a magnification changing mechanism and did not focus on long-time target tracking. The complete block diagram of the platform is shown in Figure 5. ...
We propose a microrobotic platform for single motile microorganism observation and
investigation. The platform utilizes a high-speed online vision sensor to realize real-time observation
of a microorganism under a microscopic environment with a relatively high magnification ratio.
A microfluidic chip was used to limit the vertical movement of the microorganism and reduce
the tracking system complexity. We introduce a simple image processing method, which utilizes
high-speed online vision characteristics and shows robustness against image noise to increase the
overall tracking performance with low computational time consumption. The design also considers
the future integration of a stimulation system using microtools. Successful long-time tracking of a
freely swimming microorganism inside of a microfluidic chip for more than 30 min was achieved
notwithstanding the presence of noises in the environment of the cell. The specific design of the
platform, particularly the tracking system, is described, and the performance is evaluated and
confirmed through basic experiments. The potential of the platform to apply mechanical stimulation
to a freely swimming microorganism is demonstrated by using a 50-µm-thick microtool. The proposed
platform can be used for long-term observation and to achieve different kinds of stimulations,
which can induce new behavior of the cells and lead to unprecedented discoveries in biological fields.
Based on optokinetic simulation technique, a lightweight, compactness, portable and low cost device for vestibular rehabilitation is designed and constructed. All functions are controlled remotely through a multiplatform connectivity especially developed for this application. This multiplatform not only allows for interconnection with different operating systems, but also makes it possible to control the overall clinical parameters and data acquisition and storage of the trial on each patient, thus enabling continuous monitoring of their evolution. These and other advantages are presented and demonstrated throughout this paper.
