Hybrid Segmentation Scheme for Skin Features Extraction Using Dermoscopy Images

Objective and quantitative assessment of skin conditions is essential for cosmeceutical studies and research on skin aging and skin regeneration. Various handcraft-based image processing methods have been proposed to evaluate skin conditions objectively, but they have unavoidable disadvantages when used to analyze skin features accurately. This study proposes a hybrid segmentation scheme consisting of Deeplab v3+ with an Inception-ResNet-v2 backbone, LightGBM, and morphological processing (MP) to overcome the shortcomings of handcraft-based approaches. First, we apply Deeplab v3+ with an Inception-ResNet-v2 backbone for pixel segmentation of skin wrinkles and cells. Then, LightGBM and MP are used to enhance the pixel segmentation quality. Finally, we determine several skin features based on the results of wrinkle and cell segmentation. Our proposed segmentation scheme achieved a mean accuracy of 0.854, mean of intersection over union of 0.749, and mean boundary F1 score of 0.852, which achieved 1.1%, 6.7%, and 14.8% improvement over the panoptic-based semantic segmentation method, respectively.     

Monitoring skin condition using life activities on the SNS user documents

Multimedia Tools and Applications - Social networks are not new to the IT landscape. Starting from bulletin boards and chat rooms, they have evolved to include desktop and mobile device...     

An impulse-time perturbation approach for enhancing the robustness of extra-insensitive input shapers

In this paper, perturbation-based extra-insensitive input shapers (PEI-ISs) are proposed to enhance the robustness of the input shaping technique. The extra-insensitive input shaper (EI-IS) has been known to be more robust than the so-called derivative input shapers such as ZVD, ZVDD, and ZVDDD shapers. However, the robustness of the known EI-IS is restricted by the symmetric property in the sensitivity curve. To address this, the PEI-IS is devised by multiplying a series of input shapers in the Laplace domain, of which the impulse times are slightly perturbed from those of the zero vibration (ZV) shaper. For a single-hump case, a closed-form solution to the PEI-IS is provided. For two- and three-hump cases, the approximate solutions are presented. The robustness is evaluated by simulations and assessed by means of the insensitivity. It will be shown that the proposed PEI-IS does improve the robustness and that it can be easily designed.     

An Impulse-time Perturbation Approach for a Symmetric Extra-insensitive Input Shaper

International Journal of Control, Automation and Systems - In this study, symmetric perturbation extra-insensitive input shapers (SPEI-ISs) are proposed to enhance the robustness of the...     

Electrical properties of ferroelectric SBT thin films prepared using photosensitive sol-gel solution

Self-patterning of thin films using photosensitive sol-gel solution has advantages such as simple manufacturing process compared to photoresist/dry etching process. In this study, ferroelectric Sr_0.9Bi_2.1Ta₂O₉ thin films have been prepared by spin coating method using photosensitive sol-gel solution. Strontium ethoxide, tetramethylheptanedionato bismuth and tantalum ethoxide were used as starting materials. As UV exposure time to the SBT thin film increases, the intensity of UV absorption peak of metal beta-diketonate decreases due to degradation of solubility resulted from Metal-Oxygen-Metal (M–O–M) bond formation. The solubility difference by UV irradiation on SBT thin film allows to obtain a fine patterning of thin film. The ferroelectric properties of the UV irradiated SBT thin films are superior to those of the non-UV irradiated film.     

Aeroacoustic characteristics and noise reduction of a centrifugal fan for a vacuum cleaner

The aeroacoustic characteristics of a centrifugal fan for a vacuum cleaner and its noise reduction method are studied in this paper. The major noise source of a vacuum cleaner is the centrifugal fan. The impeller of the fan rotates at over 3000 rpm, and generates very high-level noise. It was revealed that the dominant noise source is the aerodynamic interaction between the rotating impeller and stationary diffuser. The directivity of acoustic pressure showed that most of the noise propagates backward direction of the fan-motor assembly. In order to reduce the high tonal sound generated from the aerodynamic interaction, unevenly pitched impeller and diffuser, and tapered impeller designs were proposed and experiments were performed. Uneven pitch design of the impeller changes the sound quality while the overall sound power level (SPL) and the performance remains similar. The effect of the tapered design of impeller was evaluated. The trailing edge of the tapered fan is inclined. This reduces the flow interaction between the rotating impeller and the stationary diffuser because of some phase shifts. The static efficiency of the new impeller design is slightly lower than the previous design. However, the overall SPL is reduced by about 4 dB(A). The SPL of the fundamental blade passing frequency (BPF) is reduced by about 6 dB(A) and the 2^nd BPF is reduced about 20 dB(A). The vacuum cleaner with the tapered impeller design produces lower noise level than the previous one, and the strong tonal sound was dramatically reduced.     

UFC-Net with Fully-Connected Layers and Hadamard Identity Skip Connection for Image Inpainting

Image inpainting is an interesting technique in computer vision and artificial intelligence for plausibly filling in blank areas of an image by referring to their surrounding areas. Although its performance has been improved significantly using diverse convolutional neural network (CNN)-based models, these models have difficulty filling in some erased areas due to the kernel size of the CNN. If the kernel size is too narrow for the blank area, the models cannot consider the entire surrounding area, only partial areas or none at all. This issue leads to typical problems of inpainting, such as pixel reconstruction failure and unintended filling. To alleviate this, in this paper, we propose a novel inpainting model called UFC-net that reinforces two components in U-net. The first component is the latent networks in the middle of U-net to consider the entire surrounding area. The second component is the Hadamard identity skip connection to improve the attention of the inpainting model on the blank areas and reduce computational cost. We performed extensive comparisons with other inpainting models using the Places2 dataset to evaluate the effectiveness of the proposed scheme. We report some of the results.     

Effect of stents coated with a combination of sirolimus and alpha-lipoic acid in a porcine coronary restenosis model

The aim of this study was to evaluate antiproliferative sirolimus- and antioxidative alpha-lipoic acid (ALA)-eluting stents using biodegradable polymer [poly-l-lactic acid (PLA)] in a porcine coronary overstretch restenosis model. Forty coronary arteries of 20 pigs were randomized into four groups in which the coronary arteries had a bare metal stent (BMS, n = 10), ALA-eluting stent with PLA (AES, n = 10), sirolimus-eluting stent with PLA (SES, n = 10), or sirolimus- and ALA-eluting stent with PLA (SAS, n = 10). A histopathological analysis was performed 28 days after the stenting. The ALA and sirolimus released slowly over 30 days. There were no significant differences between groups in the injury or inflammation score; however, there were significant differences in the percent area of stenosis (56.2 ± 11.78 % in BMS vs. 51.5 ± 12.20 % in AES vs. 34.7 ± 7.23 % in SES vs. 28.7 ± 7.30 % in SAS, P < 0.0001) and fibrin score [1.0 (range 1.0–1.0) in BMS vs. 1.0 (range 1.0–1.0) in AES vs. 2.0 (range 2.0–2.0) in SES vs. 2.0 (range 2.0–2.0) in SAS, P < 0.0001] between the four groups. The percent area of stenosis based on micro-computed tomography corresponded with the restenosis rates based on histopathological stenosis in different proportions in the four groups (54.8 ± 7.88 % in BMS vs. 50.4 ± 14.87 % in AES vs. 34.5 ± 7.22 % in SES vs. 28.9 ± 7.22 % in SAS, P < 0.05). SAS showed a better neointimal inhibitory effect than BMS, AES, and SES at 1 month after stenting in a porcine coronary restenosis model. Therefore, SAS with PLA can be a useful drug combination for coronary stent coating to suppress neointimal hyperplasia.