Roope Raisamo

Evaluation of Gender Classification Methods with Automatically Detected and Aligned Faces

We present a systematic study on gender classification with automatically detected and aligned faces. We experimented with 120 combinations of automatic face detection, face alignment, and gender classification. One of the findings was that the automatic face alignment methods did not increase the gender classification rates. However, manual alignment increased classification rates a little, which suggests that automatic alignment would be useful when the alignment methods are further improved. We also found that the gender classification methods performed almost equally well with different input image sizes. In any case, the best classification rate was achieved with a support vector machine. A neural network and Adaboost achieved almost as good classification rates as the support vector machine and could be used in applications where classification speed is considered more important than the maximum classification accuracy.

An experimental comparison of gender classification methods

Successful face analysis requires robust methods. It has been hard to compare the methods due to different experimental setups. We carried out a comparison study for the state-of-the-art gender classification methods to find out their actual reliability. The main contributions are comprehensive and comparable classification results for the gender classification methods combined with automatic real-time face detection and, in addition, with manual face normalization. We also experimented by combining gender classifier outputs arithmetically. This lead to increased classification accuracies. Furthermore, we contribute guidelines to carry out classification experiments, knowledge on the strengths and weaknesses of the gender classification methods, and two new variants of the known methods.

Device independent text input: a rationale and an example

Individual characters and text are the main inputs in many computing devices. Currently there is a growing trend in developing small portable devices like mobile phones, personal digital assistants, GPS-navigators, and two-way pagers. Unfortunately these portable computing devices have different user interfaces and therefore the task of text input takes many forms. The user, who in the future is likely to have several of these devices, has to learn several text input methods. We argue that there is a need for a universal text input method. A method like this would work on a wide range of interface technologies and allow the user to transfer his or her writing skill without device-specific training. To show that device independent text input is possible, we present a candidate for a device independent text entry method that supports skill transfer between different devices. A limited longitudinal study was conducted to achieve a proof of concept evaluation of our Minimal Device Independent Text Input Method (MDITIM). We found MDITIM writing skill acquired with a touchpad to work almost equally well on mouse, trackball, joystick and keyboard without any additional training. Our test group reached on average 41% of their handwriting speed by the end of the tenth 30-minute training session.

Emotional and behavioral responses to haptic stimulation

A prototype of friction-based horizontally rotating fingertip stimulator was used to investigate emotional experiences and behavioral responses to haptic stimulation. The rotation style of 12 different stimuli was varied by burst length (i.e., 20, 50, 100 ms), continuity (i.e., continuous and discontinuous), and direction (e.g., forward and backward). Using these stimuli 528 stimulus pairs were presented to 12 subjects who were to distinguish if stimuli in each pair were the same or different. Then they rated the stimuli using four scales measuring the pleasantness, arousal, approachability, and dominance qualities of the 12 stimuli. The results showed that continuous forward-backward rotating stimuli were rated as significantly more unpleasant, arousing, avoidable, and dominating than other types of stimulations (e.g., discontinuous forward rotation). The reaction times to these stimuli were significantly faster than reaction times to discontinuous forward and backward rotating stimuli. The results clearly suggest that even simple haptic stimulation can carry emotional information. The results can be utilized when making use of haptics in human-technology interaction.

Methods for Presenting Braille Characters on a Mobile Device with a Touchscreen and Tactile Feedback

Three novel interaction methods were designed for reading six-dot Braille characters from the touchscreen of a mobile device. A prototype device with a piezoelectric actuator embedded under the touchscreen was used to create tactile feedback. The three interaction methods, scan, sweep, and rhythm, enabled users to read Braille characters one at a time either by exploring the characters dot by dot or by sensing a rhythmic pattern presented on the screen. The methods were tested with five blind Braille readers as a proof of concept. The results of the first experiment showed that all three methods can be used to convey information as the participants could accurately (91-97 percent) recognize individual characters. In the second experiment the presentation rate of the most efficient and preferred method, the rhythm, was varied. A mean recognition accuracy of 70 percent was found when the speed of presenting a single character was nearly doubled from the first experiment. The results showed that temporal tactile feedback and Braille coding can be used to transmit single-character information while further studies are still needed to evaluate the presentation of serial information, i.e., multiple Braille characters.

Appropriateness of foot interaction for non-accurate spatial tasks

This paper describes alternative methods for manipulating graphical user interfaces with a foot. Feet are used in many real world tasks together with the rest of the body, but in computer environments they are almost completely put aside as an interaction possibility. One of the major problems in choosing input methods for different tasks in user interfaces is determining what kind of method is appropriate for a certain task. Feet could easily be used as a supportive input method in interaction with computers together with the traditional mouse. In this paper, we discuss the possibility of using foot input in different non-accurate spatial tasks, and the efficiency and usability experience the users have of foot interaction compared with a traditional hand-based interface with the same input device. The aim is to find out how well foot interaction suits for non-accurate spatial tasks.

Quikwriting as a multi-device text entry method

Quikwriting is a previously published technique for entering text into computers using a stylus. We report results of a longitudinal study on user performance with it. In addition to the original stylus-based usage mode we designed modes for joystick and keyboard thus making Quikwriting compatible with a wide range of computing devices. Twelve participants used the stylus and joystick modes in 20 sessions for a total of ten hours. By the end of the experiment their text entry rate was 16 wpm in the stylus mode and 13 wpm in the joystick mode. At the end we conducted a test to verify that Quikwriting skill transfers to the keyboard mode. Text entry rate for the first five minutes of use in the keyboard mode was 6 wpm. In summary, the stylus mode was not particularly fast, but we found Quikwriting suitable for multi-device use.

Gaze gestures and haptic feedback in mobile devices

Anticipating the emergence of gaze tracking capable mobile devices, we are investigating the use of gaze as an input modality in handheld mobile devices. We conducted a study of combining gaze gestures with vibrotactile feedback. Gaze gestures were used as an input method in a mobile device and vibrotactile feedback as a new alternative way to give confirmation of interaction events. Our results show that vibrotactile feedback significantly improved the use of gaze gestures. The tasks were completed faster and rated easier and more comfortable when vibrotactile feedback was provided.

Design and evaluation of a tactile memory game for visually impaired children

Visually impaired people have a lack of proper user interfaces to allow them to easily make use of modern technology. This problem may be solved with multimodal user interfaces that should be designed taking into account the type and degree of disability. The purpose of the study presented in this article was to create usable games for visually impaired children making use of low-cost vibro-tactile devices in multimodal applications. A tactile memory game using multimodal navigation support with high-contrast visual feedback and audio cues was implemented. The game was designed to be played with a tactile gamepad. Different vibrations were to be remembered instead of sounds or embossed pictures that are common in memory games for blind children. The usability and playability of the game was tested with a group of seven 12-13-year-old visually impaired children. The results showed that the game design was successful and a tactile gamepad was usable. The game got a positive response from the focus group.

Touch gestures in communicating emotional intention via vibrotactile stimulation

Remote communication between people typically relies on audio and vision although current mobile devices are increasingly based on detecting different touch gestures such as swiping. These gestures could be adapted to interpersonal communication by using tactile technology capable of producing touch stimulation to a users hand. It has been suggested that such mediated social touch would allow for new forms of emotional communication. The aim was to study whether vibrotactile stimulation that imitates human touch can convey intended emotions from one person to another. For this purpose, devices were used that converted touch gestures of squeeze and finger touch to vibrotactile stimulation. When one user squeezed his device or touched it with finger(s), another user felt corresponding vibrotactile stimulation on her device via four vibrating actuators. In an experiment, participant dyads comprising a sender and receiver were to communicate variations in the affective dimensions of valence and arousal using the devices. The senders task was to create stimulation that would convey unpleasant, pleasant, relaxed, or aroused emotional intention to the receiver. Both the sender and receiver rated the stimulation using scales for valence and arousal so that the match between senders intended emotions and receivers interpretations could be measured. The results showed that squeeze was better at communicating unpleasant and aroused emotional intention, while finger touch was better at communicating pleasant and relaxed emotional intention. The results can be used in developing technology that enables people to communicate via touch by choosing touch gesture that matches the desired emotion.