• The Transactions of the Korea Information Processing Society
• /
• v.13 no.4
• /
• pp.189-198
• /
• 2024

Lipreading is one of the important parts of speech recognition, and several studies have been conducted to improve the performance of lipreading in lipreading systems for speech recognition. Recent studies have used method to modify the model architecture of lipreading system to improve recognition performance. Unlike previous research that improve recognition performance by modifying model architecture, we aim to improve recognition performance without any change in model architecture. In order to improve the recognition performance without modifying the model architecture, we refer to the cues used in human lipreading and set other regions such as chin and cheeks as regions of interest along with the lip region, which is the existing region of interest of lipreading systems, and compare the recognition rate of each region of interest to propose the highest performing region of interest In addition, assuming that the difference in normalization results caused by the difference in interpolation method during the process of normalizing the size of the region of interest affects the recognition performance, we interpolate the same region of interest using nearest neighbor interpolation, bilinear interpolation, and bicubic interpolation, and compare the recognition rate of each interpolation method to propose the best performing interpolation method. Each region of interest was detected by training an object detection neural network, and dynamic time warping templates were generated by normalizing each region of interest, extracting and combining features, and mapping the dimensionality reduction of the combined features into a low-dimensional space. The recognition rate was evaluated by comparing the distance between the generated dynamic time warping templates and the data mapped to the low-dimensional space. In the comparison of regions of interest, the result of the region of interest containing only the lip region showed an average recognition rate of 97.36%, which is 3.44% higher than the average recognition rate of 93.92% in the previous study, and in the comparison of interpolation methods, the bilinear interpolation method performed 97.36%, which is 14.65% higher than the nearest neighbor interpolation method and 5.55% higher than the bicubic interpolation method. The code used in this study can be found a https://github.com/haraisi2/Lipreading-Systems.

• Journal of Korean Elementary Science Education
• /
• v.43 no.3
• /
• pp.415-432
• /
• 2024

This study developed and applied the teacher education model and its principles for science classes using Virtual and Augmented Reality (VR/AR) content and analyzed preservice elementary teachers' feedback on the teacher education model and the changes in their perceptions as to the use of VR/AR content. First, existing Technological Pedagogical Content Knowledge (TPACK) teacher education models and prior studies on the use of the VR/AR contents were reviewed to derive the teacher education model to cultivate the VR/AR-TPACK and set the key principles for each of its stages. The developed teacher education model has five stages: exploration, mapping, collaborative design, practice, and reflection. Second, to examine the appropriateness of the model's five stages and principles, we applied it within the regular course of instruction at the university of education, which was attended by 25 preservice elementary teachers. This study collected data from surveys on the perception of the usage of VR/AR contents before and after the course, as well as the group lesson plans prepared by the preservice teachers, and their feedback on the teacher education model. The feedback on the teacher education model and the survey conducted by the preservice teachers before and after the course were analyzed through open coding and categorization. As a result, most preservice teachers expressed positive opinions about the activities and experiences at each stage of the implementation of the teacher education model. Perceptions related to the usage of the VR/AR content changed in three aspects: first, the vague positive perception of the VR/AR content has changed to a positive perception based on specific educational affordance. Second, they recognized the need for preparedness by anticipating potential problems associated with the use of the VR/AR content. Third, they came to view the VR/AR contents as a useful instructional resource that the teachers could use. Based on these results, we discussed the implications for the VR/AR-TPACK teacher education model and assessed the limitations of the research.

• Nuclear Medicine and Molecular Imaging
• /
• v.41 no.1
• /
• pp.9-15
• /
• 2007

Purpose: Repetitive transcranial magnetic stimulation (rTMS) has recently been clinically applied in the treatment of drug resistant depressed patients. There are mixed findings about the efficacy of rTMS on depression. Furthermore, the influence of rTMS on the physiology of the brain is not clear. We prospectively evaluated changes of regional cerebral blood flow (rCBF) between pre- and post-rTMS treatment in patients with drug resistant depression. Materials and Methods: Twelve patients with drug-resistant depression (7 male, 5 female; age range: $19{\sim}52$ years; mean age: $29.3{\pm}9.3$ years) were given rTMS on right prefrontal lobe with low frequency (1 Hz) and on left prefrontal lobe with high frequency (20 Hz), with 20-minute-duration each day for 3 weeks. Tc-99m ECD brain perfusion SPECT was obtained before and after rTMS treatment. The changes of cerebral perfusion were analyzed using statistical parametric mapping (SPM; t=3.14, uncorrected p<0.01, voxel=100). Results: Following areas showed significant increase in rCBF after 3 weeks rTMS treatment: the cingulate gyrus, fusiform gyrus of right temporal lobe, precuneus, and left lateral globus pallidus. Significant decrement was noted in: the precental and middle frontal gyrus of right frontal lobe, and fusiform gyrus of left occipital lobe. Conclusion: Low-frequency rTMS on the right prefrontal cortex and high-frequency rTMS on the left prefrontal cortex for 3 weeks as an add-on regimen have increased and decreased rCBF in the specific brain regions in drug-resistant depressed patients. Further analyses correlating clinical characteristics and treatment paradigm with functional imaging data may be helpful in clarifying the pathophysiology of drug-resistant depressed patients.