MedOptNet: Meta-Learning Framework for Few-shot Medical Image Classification.

In the medical research domain, limited data and high annotation costs have made efficient classification under few-shot conditions a popular research area. This paper proposes a meta-learning framework, termed MedOptNet, for few-shot medical image classification. The framework enables the use of various high-performance convex optimization models as classifiers, such as multi-class kernel support vector machines, ridge regression, and other models. End-to-end training is then implemented using dual problems and differentiation in the paper. Additionally, various regularization techniques are employed to enhance the model's generalization capabilities. Experiments on the BreakHis, ISIC2018, and Pap smear medical few-shot datasets demonstrate that the MedOptNet framework outperforms benchmark models. Moreover, the model training time is also compared to prove its effectiveness in the paper, and an ablation study is conducted to validate the effectiveness of each module.

Shooting Utility Maximization in UAV-Assisted Wireless Camera Sensor Networks.

Recently, wireless camera sensor networks (WCSNs) have entered an era of rapid development, and WCSNs assisted by unmanned aerial vehicles (UAVs) are capable of providing enhanced flexibility, robustness and efficiency when executing missions such as shooting targets. Existing research has mainly focused on back-end image processing to improve the quality of captured images, but it has neglected the question of attaining quality images on the front-end, which is significantly influenced by the location and hovering time of the UAV. Therefore, in this paper, we conceive a novel shooting utility model to quantify shooting quality, which is maximized by simultaneously considering the UAV's trajectory planning, hovering time and shooting point selection. To expound further, we prove the submodularity of the utility function, whereby the original problem can be expressed as a submodular maximization problem with path constraints, and we propose a utility-cost ratio (UCR) algorithm to maximize shooting utility through two-level optimization. Then, by using the relaxation of the cost function, we analyze the gap between the proposed algorithm and the optimal algorithm (OPT) and prove that the UCR algorithm has a bi-criterion approximation ratio of 1-1/e/2. Simulation results show that the algorithm outperforms both the random algorithm (RAN) and the maximum shooting utility point selection algorithm (MSU) in terms of shooting utility and time utilization efficiency, improving shooting utility by 51% and 21% compared to the RAN and MSU algorithms, respectively, and achieving at least 88.2% of the OPT algorithm in terms of time utilization efficiency.

ANN Assisted-IoT Enabled COVID-19 Patient Monitoring.

COVID-19 is an extremely dangerous disease because of its highly infectious nature. In order to provide a quick and immediate identification of infection, a proper and immediate clinical support is needed. Researchers have proposed various Machine Learning and smart IoT based schemes for categorizing the COVID-19 patients. Artificial Neural Networks (ANN) that are inspired by the biological concept of neurons are generally used in various applications including healthcare systems. The ANN scheme provides a viable solution in the decision making process for managing the healthcare information. This manuscript endeavours to illustrate the applicability and suitability of ANN by categorizing the status of COVID-19 patients' health into infected (IN), uninfected (UI), exposed (EP) and susceptible (ST). In order to do so, Bayesian and back propagation algorithms have been used to generate the results. Further, viterbi algorithm is used to improve the accuracy of the proposed system. The proposed mechanism is validated over various accuracy and classification parameters against conventional Random Tree (RT), Fuzzy C Means (FCM) and REPTree (RPT) methods.

Efficient light harvesting with a nanostructured organic electron-transporting layer in perovskite solar cells.

Nanostructures have proved to be advantageous in light harvesting, improving the power conversion efficiency (PCE) of photovoltaic devices. However, the reported light-harvesting nanostructures all require extra processing beyond that for device fabrication, with multiple steps for nano-patterned structures or plasmonic nano-particles. Here we synthesized a conjugated polymer PFPDI which could simply form a nanostructured film on perovskite by spin coating. PFPDI possesses a perylene diimide-based backbone and phosphite ester pendants, which makes it a robust electron-transporting material (ETM) in perovskite solar cells. Furthermore, the perovskite solar cells with PFPDI as the electron-transporting layer (ETL) exhibited a higher light-harvesting efficiency compared to the standard phenyl-C61-butyric acid methyl ester (PCBM) devices. The JSC of the PFPDI device was enhanced from 19.71 mA cm-2 to 23.43 mA cm-2. SEM images and reflectance spectra confirmed that the PFPDI formed ditch-like nanostructures on perovskite film and induced a better light-harvesting efficiency. Further research indicated that the interaction of P[double bond, length as m-dash]O with Pb was essential to the formation of the nanostructures of PFPDI on perovskite. Therefore, our work not only offers an efficient organic ETM, but also opens up new horizons for simply constructing nanostructures with light-harvesting ability in photovoltaic devices.

High-Performance Polymer Solar Cells with Zinc Sulfide-Phenanthroline Derivatives as the Hybrid Cathode Interlayers.

Environmentally benign hybrid interlayers are prepared by modifying the zinc sulfide (ZnS) with phenanthroline/derivatives and utilized in inverted polymer solar cells (PSCs). Performances of the inverted PSCs are improved enormously by incorporating these hybrid interlayers, as which can effectively improve the energy level alignment, electron mobility, surface morphology, and interfacial contact. Greatly improved power conversion efficiencies (PCEs) of 7.79%, 8.00%, 7.47%, and 7.56% are achieved with these hybrid interlayers ZnS-BCP, ZnS-Bphen, ZnS-Mphen, and ZnS-Phen, respectively, compared to the PCE of 2.99% of the reference ZnS-based device, based on PTB7:PC71BM active layer. Our results demonstrate that hybrid interfacial materials comprising inorganic and organic semiconductor possess promising potential to improve the performance of organic electronic devices, and set an example to develop this novel class of interfacial materials for electronic devices.

Side-chain engineering of benzodithiophene-fluorinated quinoxaline low-band-gap co-polymers for high-performance polymer solar cells.

A new series of donor-acceptor co-polymers based on benzodithiophene and quinoxaline with various side chains have been developed for polymer solar cells. The effect of the degree of branching and dimensionality of the side chains were systematically investigated on the thermal stability, optical absorption, energy levels, molecular packing, and photovoltaic performance of the resulting co-polymers. The results indicated that the linear and 2D conjugated side chains improved the thermal stabilities and optical absorptions. The introduction of alkylthienyl side chains could efficiently lower the energy levels compared with the alkoxyl-substituted analogues, and the branched alkoxyl side chains could deepen the HOMO levels relative to the linear alkoxyl chains. The branched alkoxyl groups induced better lamellar-like ordering, but poorer face-to-face packing behavior. The 2D conjugated side chains had a negative influence on the crystalline properties of the co-polymers. The performance of the devices indicated that the branched alkoxyl side chains improved the Voc, but decreased the Jsc and fill factor (FF). However, the 2D conjugated side chains would increase the Voc, Jsc, and FF simultaneously. For the first time, our work provides insight into molecular design strategies through side-chain engineering to achieve efficient polymer solar cells by considering both the degree of branching and dimensionality.

Improving efficiency by hybrid TiO(2) nanorods with 1,10-phenanthroline as a cathode buffer layer for inverted organic solar cells.

We reported a significant improvement in the efficiency of organic solar cells by introducing hybrid TiO2:1,10-phenanthroline as a cathode buffer layer. The devices based on polymer thieno[3,4-b]thiophene/benzodithiophene:[6,6]-phenyl C71-butyric acid methyl ester (PTB7:PC71BM) with hybrid buffer layer exhibited an average power conversion efficiency (PCE) as high as 8.02%, accounting for 20.8% enhancement compared with the TiO2 based devices. The cathode modification function of this hybrid material could also be extended to the poly(3-hexylthiophene):[6,6]-phenyl-C61-butyric acid methyl ester (P3HT:PC61BM) system. We anticipate that this study will stimulate further research on hybrid materials to achieve more efficient charge collection and device performance.

Solution-processed hybrid cathode interlayer for inverted organic solar cells.

A novel hybrid material CdS/2,9-Dimethyl-4,7-diphenyl-1,10-phenanthroline (CdS·BCP) was prepared from the decomposition of its organic soluble precursor complex Cd(S2COEt)2·(BCP) by low-temperature treatment. CdS·BCP, which integrated the favorable properties of solvent durability, and high electron mobility of CdS as well as the good hole blocking property of BCP, was designed and developed as the interface modification material to improve electron collection in bulk heterojunction organic solar cells (OSCs). The inverted OSCs with CdS·BCP as buffer layer on ITO showed improved efficiency compared with the pure CdS or BCP. Devices with CdS·BCP as interlayer exhibited excellent stability, only 14.19% decay of power conversion efficiencies (PCEs) was observed (from 7.47% to 6.41%) after stored in glovebox for 3264 h (136 days). Our results demonstrate promising potentials of hybrid materials as the interface modification layers in OSCs, and provide new insights for the development of new interface modification materials in the future.

Solution-processed MoS(x) as an efficient anode buffer layer in organic solar cells.

We reported a facile solution-processed method to fabricate a MoSx anode buffer layer through thermal decomposition of (NH4)2MoS4. Organic solar cells (OSCs) based on in situ growth MoSx as the anode buffer layer showed impressive improvements, and the power conversion efficiency was higher than that of conventional PEDOT:PSS-based device. The MoSx films obtained at different temperatures and the corresponding device performance were systematically studied. The results indicated that both MoS3 and MoS2 were beneficial to the device performance. MoS3 could result in higher Voc, while MoS2 could lead to higher Jsc. Our results proved that, apart from MoO3, molybdenum sulfides and Mo(4+) were also promising candidates for the anode buffer materials in OSCs.

[Application of bi-pedicel and bi-muscle flaps for the expanded partial laryngectomy in the patients with T3T4 laryngeal cancer].

OBJECTIVE: To evaluate the application of bi-pedicel and bi-muscle flaps for the expanded partial laryngectomy in the patients with T3T4 larynx cancer. METHODS: From 1990 to 1997, 386 patients of laryngeal cancer were treated by expanded partial laryngectomy and repaired with unilateral or bilateral bi-pedicel and bi-muscle flaps. 322 patients were male and 64 were female. The age ranged from 30 to 84 years old. All the patients were in the stage T3 or T4 of cancer. The operation included expanded horizontal supraglottic laryngectomy, vertical partial laryngectomy, 3/4 laryngectomy and sub-total laryngectomy. RESULTS: After operation, the speech function was still good in all cases. The swallow function recovered in 8 to 12 days post-operation. The extraction rate of tracheal tube was 50.3% (194/386), and the 5-years survival rate was 55.2% (213/386) while the local recurrence rate was 16.6% (64/386). CONCLUSION: With the application of bi-pedicel and bi-muscle flaps, the patients, if treated by total laryngectomy as traditional technique, could be treated by partial laryngectomy. This method could expand the indication of partial laryngectomy and improve the life quality and survival rate.