Unifying Convolution and Transformer Decoder for Textile Fiber Identification

At present,convolutional neural networks(CNNs)and transformers surpass humans in many situations(such as face recognition and object classification),but do not work well in identifying fibers in textile surface images.Hence,this paper proposes an architecture named FiberCT which takes advantages of the feature extraction capability of CNNs and the long-range modeling capability of transformer decoders to adaptively extract multiple types of fiber features.Firstly,the convolution module extracts fiber features from the input textile surface images.Secondly,these features are sent into the transformer decoder module where label embeddings are compared with the features of each type of fibers through multi-head cross-attention and the desired features are pooled adaptively.Finally,an asymmetric loss further purifies the extracted fiber representations.Experiments show that FiberCT can more effectively extract the representations of various types of fibers and improve fiber identification accuracy than state-of-the-art multi-label classification approaches.

Evaluation of pure-tone audiometric protocols in vestibular schwannoma screening

The objective was to evaluate the pure-tone audiogram-based screening protocols in VS diagnostics.We retrospectively analyzed presenting symptoms,pure tone audiometry and MRI finding from 246 VS patients and 442 controls were collected to test screening protocols(AAO-HNS,AMCLASS-A/B,Charing Cross,Cueva,DOH,Nashville,Oxford,Rule3000,Schlauch,Seattle,Sunderland)for sensitivity and specificity.Results were pooled with data from five other studies,and analysis of sensitivity,specificity and positive likelihood ratio(LR+)for each protocol was performed.Our results show that protocols with significantly higher sensitivity(AMCLASS-A/B,Nashville)show also significantly lowest specificity,and tend to have low association(positive likelihood ratio,LR+)to the VS.The highest LR+was found for protocols AAO-HNS,Rule3000 and Seattle.In conclusions,knowing their properties,screening protocols are simple decision-making tools in VS diagnostic.To use the advantage of the highest sensitivity,protocols AMCLASS-A+B or Nashville can be of choice.For more reasonable approach,applying the protocols with high LR+(AAO-HNS,Rule3000,Seattle)may reduce the overall number of MRI scans at expense of only few primarily undiagnosed VS.

Asymmetric twin-field quantum key distribution with both statistical and intensity fluctuations

Twin-field quantum key distribution(TF-QKD)is a disruptive innovation which is able to overcome the rate-distance limit of QKD without trusted relays.Since the proposal of the first TF-QKD protocol,theoretical and experimental breakthroughs have been made to enhance its ability.However,there still exist some practical issues waiting for settlement.In this paper,we examine the performances of asymmetric TF-QKD protocol with unstable light sources and limited data sizes.The statistical fluctuations of the parameters are estimated employing Azuma’s inequality.Through numerical simulations,we compare the secret key rates of the asymmetric TF-QKD protocol with different data sizes and variant intensity fluctuation magnitudes.Our results demonstrate that both statistical and intensity fluctuations have significant impacts on the performance of asymmetric TF-QKD.