No abstract

In a scenario of worldwide honey bee decline, assessing colony strength is becoming increasingly important for sustainable beekeeping. Temporal counts of number of comb cells with brood and food reserves offers researchers data for multiple applications, such as modelling colony dynamics, and beekeepers information on colony strength, an indicator of colony health and honey yield. Counting cells manually in comb images is labour intensive, tedious, and prone to error. Herein, we developed a free software, named DeepBee©, capable of automatically detecting cells in comb images and classifying their contents into seven classes. By distinguishing cells occupied by eggs, larvae, capped brood, pollen, nectar, honey, and other, DeepBee© allows an unprecedented level of accuracy in cell classification. Using Circle Hough Transform and the semantic segmentation technique, we obtained a cell detection rate of 98.7%, which is 16.2% higher than the best result found in the literature. For classification of comb cells, we trained and evaluated thirteen different convolutional neural network (CNN) architectures, including: DenseNet (121, 169 and 201); InceptionResNetV2; InceptionV3; MobileNet; MobileNetV2; NasNet; NasNetMobile; ResNet50; VGG (16 and 19) and Xception. MobileNet revealed to be the best compromise between training cost, with~9 s for processing all cells in a comb image, and accuracy, with an F1-Score of 94.3%. We show the technical details to build a complete pipeline for classifying and counting comb cells and we made the CNN models, source code, and datasets publicly available. With this effort, we hope to have expanded the frontier of apicultural precision analysis by providing a tool with high performance and source codes to foster improvement by third parties (https://github.com/AvsThiago/DeepBeesource).

Respiratory insufficiency is a symptom that requires hospitalization. This work investigates whether it is possible to detect this condition by analyzing patient's speech samples; the analysis was performed on data collected during the first wave of the COVID-19 pandemic in 2020, and thus limited to respiratory insufficiency in COVID-19 patients. For that, a dataset was created consisting of speech emissions of both COVID-19 patients affected by respiratory insufficiency and a control group. This dataset was used to build a Convolution Neural Network to detect respiratory insufficiency using speech emission MFCC representations. Methodologically, dealing with background noise was a challenge, so we also collected background noise from COVID-19 wards where patients were located. Due to the difficulty in filtering noise without eliminating crucial information, noise samples were injected in the control group data to prevent bias. Moreover, we investigated (i) two approaches to address the duration variance of audios, and (ii) the ideal number of noise samples to inject in both patients and the control group to prevent bias and overfitting. The techniques developed reached 91.66% accuracy. Thus we validated the project's Leading Hypothesis, namely that it is possible to detect respiratory insufficiency in speech utterances, under real-life environmental conditions; we believe our results justify further enquiries into the use of automated speech analysis to support health professionals in triage procedures.

No abstract

In this paper we describe a web content adaptation tool for assisting low-literacy readers to access online information. The"Educational FACILITA" tool provides innovative features and the design of more intuitive interaction models. Especially, we propose an interaction model and web application that explore the Natural Language Processing tasks of lexical elaboration and named entity labeling for improving web accessibility.

In this work, we propose several techniques to address data scarceness in ComParE 2021 COVID-19 identification tasks for the application of deep models such as Convolutional Neural Networks. Data is initially preprocessed into spectrogram or MFCC-gram formats. After preprocessing, we combine three different data augmentation techniques to be applied in model training. Then we employ transfer learning techniques from pretrained audio neural networks. Those techniques are applied to several distinct neural architectures. For COVID-19 identification in speech segments, we obtained competitive results. On the other hand, in the identification task based on cough data, we succeeded in producing a noticeable improvement on existing baselines, reaching 75.9% unweighted average recall (UAR).

No abstract

YourTTS brings the power of a multilingual approach to the task of zero-shot multi-speaker TTS. Our method builds upon the VITS model and adds several novel modifications for zeroshot multi-speaker and multilingual training. We achieved stateof-the-art (SOTA) results in zero-shot multi-speaker TTS and results comparable to SOTA in zero-shot voice conversion on the VCTK dataset. Additionally, our approach achieves promising results in a target language with a single-speaker dataset, opening possibilities for zero-shot multi-speaker TTS and zeroshot voice conversion systems in low-resource languages. Finally, it is possible to fine-tune the YourTTS model with less than 1 minute of speech and achieve state-of-the-art results in voice similarity and with reasonable quality. This is important to allow synthesis for speakers with a very different voice or recording characteristics from those seen during training.