2D Human Parsing with Deep Skin Model and Part-Based Model Inference

Author(s):  
Tao Xu ◽  
Zhiquan Feng ◽  
Likai Dong ◽  
Xiaohui Yang
Keyword(s):  
2020 ◽  
Vol 16 (6) ◽  
pp. 722-737
Author(s):  
Cigdem Yengin ◽  
Emrah Kilinc ◽  
Fatma Gulay Der ◽  
Mehmet Can Sezgin ◽  
Ilayda Alcin

Background: Reverse İontophoresis (RI) is one of the promising non-invasive technologies. It relies on the transition of low magnitude current through the skin and thus glucose measurement becomes possible as it is extracted from the surface during this porter current flow. Objective: This paper deals with the development and optimization of an RI determination method for glucose. CE dialysis membrane based artificial skin model was developed and the dependence of RI extraction on various experimental parameters was investigated. Method: Dependence of RI extraction performance on noble electrodes (platinum, silver, palladium, ruthenium, rhodium) was checked with CA, CV and DPV, in a wide pH and ionic strength range. Optimizations on inter-electrode distance, potential type and magnitude, extraction time, gel type, membrane MWCO, usage frequency, pretreatment, artificial body fluids were performed. Results: According to the optimized results, the inter-electrode distance was 7.0 mm and silver was the optimum noble metal. Optimum pH and ionic strength were achieved with 0.05M PBS at pH 7.4. Higher glucose yields were obtained with DPV, while CA and CV achieved almost the same levels. During CA, +0.5V achieved the highest glucose yield and higher potential even caused a decrease. Glucose levels could be monitored for 24 hours. CMC gel was the optimum collection media. Pretreated CE membrane with 12kD MWCO was the artificial skin model. Pretreatment affected the yields while its condition caused no significant difference. Except PBS solution (simulated as artificial plasma), among the various artificial simulated body fluids, intestinal juice formulation (AI) and urine formulation U2 were the optimum extraction media, respectively. Conclusion: In this study, various experimental parameters (pretereatment procedure, type and MWCO values of membranes, inter-electrode distance, electrode material, extraction medium solvents, ionic strength and pH, collection medium gel type, extraction potential type and magnitude, extraction time and etc) were optimized for the non-invasive RI determination of glucose in a CE dialysis membrane-based artificial skin model and various simulated artificial body fluids.


Author(s):  
Ryohei Hirose ◽  
Risa Bandou ◽  
Hiroshi Ikegaya ◽  
Naoto Watanabe ◽  
Takuma Yoshida ◽  
...  

1992 ◽  
Vol 8 (4) ◽  
pp. 267-276 ◽  
Author(s):  
Sandra R. Slivka

Author(s):  
Zhuliang Yao ◽  
Shijie Cao ◽  
Wencong Xiao ◽  
Chen Zhang ◽  
Lanshun Nie

In trained deep neural networks, unstructured pruning can reduce redundant weights to lower storage cost. However, it requires the customization of hardwares to speed up practical inference. Another trend accelerates sparse model inference on general-purpose hardwares by adopting coarse-grained sparsity to prune or regularize consecutive weights for efficient computation. But this method often sacrifices model accuracy. In this paper, we propose a novel fine-grained sparsity approach, Balanced Sparsity, to achieve high model accuracy with commercial hardwares efficiently. Our approach adapts to high parallelism property of GPU, showing incredible potential for sparsity in the widely deployment of deep learning services. Experiment results show that Balanced Sparsity achieves up to 3.1x practical speedup for model inference on GPU, while retains the same high model accuracy as finegrained sparsity.


2021 ◽  
Vol 158 ◽  
pp. 266-272
Author(s):  
Loris Busch ◽  
Yasemin Keziban ◽  
Lars Dähne ◽  
Cornelia M. Keck ◽  
Martina C. Meinke ◽  
...  

2003 ◽  
Vol 22 (5) ◽  
pp. 255-261 ◽  
Author(s):  
R P Chilcott ◽  
C H Dalton ◽  
I Hill ◽  
C M Davidson ◽  
K L Blohm ◽  
...  

Nerve agents are a class of organophosphorus chemicals that inhibit certain cholinesterase enzymes (ChE). If untreated, percutaneous exposure to nerve agents, such as VX (O-ethyl-S-[2(diisopropylamino)ethyl] methylphosphonothioate) can cause paralysis, apnoea and death. Much of the information concerning the percutaneous absorption and subsequent toxicity of nerve agents has been obtained using various rodent models. However, the most relevant ‘skin model’ is arguably the pig. Therefore, the purpose of this study was to examine the clinical manifestations of VX intoxication in the domestic white pig following a 2 LD50 (120 mg/kg) percutaneous challenge. There was a consistent onset of signs (where present) in each animal: mastication was followed by miosis, salivation, fasciculations and apnoea. Whilst ChE activity did not correlate with the onset of signs, there was a qualitative relationship in that mastication preceded substantial ChE inhibition, miosis lagged behind the linear decrease in acetylcholinesterase (AChE) activity and fasciculations and apnoea occurred after maximum ChE inhibition had been attained (5 / 10% of normal). These observations may be of use for the triage of patients exposed to VX. In comparison with similar studies with GD, VX did not affect glucose utilization. However, VX was similar to GD in that it caused a mild hyperkalaemia and hyperphosphataemia, although the significance of this observation was not clear. There was substantial lateral diffusion of the initial droplet of VX over the application site, indicating that, when decontaminating exposed skin, attention should also be directed to areas peripheral to the original site of exposure.


2014 ◽  
Vol 227 (2) ◽  
pp. 139-149 ◽  
Author(s):  
Thalita B. Zanoni ◽  
Manoela Tiago ◽  
Fernanda Faião-Flores ◽  
Silvia B. de Moraes Barros ◽  
Aalt Bast ◽  
...  

2015 ◽  
Vol 18 (4) ◽  
pp. 578 ◽  
Author(s):  
Bradley S Simpson ◽  
Xianling Luo ◽  
Jiping Wang ◽  
Yunmei Song ◽  
David Claudie ◽  
...  

Purpose: We have previously reported that the Australian Northern Kaanju (Kuuku I’yu) medicinal plant Dodonaea polyandra has anti-inflammatory activity. This is attributed largely to the presence of clerodane diterpenoids contained within the leaf resin. We envisaged developing a topical preparation to treat indications relating to skin inflammation. However, it was unknown whether the resin could be incorporated into a suitable dosage form while retaining the therapeutic value demonstrated in previous work. Therefore, the following study was undertaken to assess parameters of safety and efficacy for a prototype formulation containing the leaf resin extracted from D. polyandra. Methods: Using the assessment criteria of optimum appearance, tactile feeling, spreadability and odour, 78 different formulations were developed. Formulation stability was assessed using a centrifugal test with preparations displaying phase separation further modified or re-formulated. A prototype formulation containing 5% w/w plant resin was selected and subjected to in vitro release studies. This was quantified through HPLC analysis using two major bioactive diterpenoids as reference. The prototype formulation was tested for efficacy in a TPA-induced acute murine skin inflammation model as well as a 3D human skin model for irritancy/toxicity (Epiderm™). Results: The prototype resin cream was a chartreuse-coloured homogenous semisolid preparation that was readily spreadable upon contact with skin with no sensation of tackiness, residual greasiness, or irritation. The optimized cream showed no phase separation after 30 min centrifugation at 825 g. In the TPA-induced inflammation model, the resin formulation significantly reduced ear thickness and interleukin-1 beta levels in mouse ear tissue. The 5% w/w resin cream formulation showed no irritancy in a 3D human skin model. Conclusions: Our results demonstrate that bioactive resin from D. polyandra can be formulated into a stable and non-irritant semi-solid dosage form and reduce parameters of acute skin inflammation in vivo. This article is open to POST-PUBLICATION REVIEW. Registered readers (see “For Readers”) may comment by clicking on ABSTRACT on the issue’s contents page.


1988 ◽  
Vol 3 (5) ◽  
pp. 265-271 ◽  
Author(s):  
Koji Komatsu

Sign in / Sign up

Export Citation Format

Share Document