Samsung (Brazil)

Pulse wave velocity (PWV) has been recognised as the gold standard for assessing arterial stiffness and a relevant indicator in diagnosing cardiovascular disease. Conventional approaches can be affected by factors such as the size of the probe, its positioning on the skin with the appropriate angle and magnitude of the incident force, or influenced by optical properties. Aiming at improving the assessment of PWV parameter, an important cardiovascular risk marker, the present study introduces a new arterial pulse wave measurement technique based on measurements of the impedance phase characteristics of giant magneto-impedance (GMI) sensors submitted to slight magnetic field variations caused by the displacement of a small magnetic marker placed on the patient's skin, whose movement is coordinated by the local pressure wave. The proposed method eliminates the necessity of using probes with mechanical amplification, enhancing spatial resolution and usability in hard-to-reach anatomical regions through a contactless device unaffected by optical parameters. The obtained experimental results indicate the potential of the developed measurement system in measuring arterial pulse waveform and PWV.

Mutation testing is widely recognized as one of the most effective criteria for software testing to uncover defects. However, there is an urgent need for more significant efforts to promote learning and understanding of this criterion. Currently, teaching support mechanisms in software testing tend to prioritize other criteria over mutation testing, exacerbating the gap. In this paper, we introduce MUT-STT, an open educational resource specifically designed to address this gap and facilitate the dissemination of critical concepts and practical aspects of mutation testing. MUT-STT focuses on the intrinsic concepts of the mutation testing criterion, providing comprehensive support for learners. It incorporates review exercises, offers tool suggestions for test automation, and provides clear instructions on applying the mutation test criterion using Python. An evaluation was conducted to ensure the educational resource’s content was complete and correct. In this evaluation, experts in mutation testing carefully examined MUT-STT, identifying and addressing any conceptual or descriptive defects that may have persisted.

The recognition of human activities (HAR) using wearable device data, such as smartwatches, has gained significant attention in the field of computer science due to its potential to provide insights into individuals’ daily activities. This article aims to conduct a comparative study of deep learning techniques for recognizing activities of daily living (ADL). A mapping of HAR techniques was performed, and three techniques were selected for evaluation, along with a dataset. Experiments were conducted using the selected techniques to assess their performance in ADL recognition, employing standardized evaluation metrics, such as accuracy, precision, recall, and F1-score. Among the evaluated techniques, the DeepConvLSTM architecture, consisting of recurrent convolutional layers and a single LSTM layer, achieved the most promising results. These findings suggest that software applications utilizing this architecture can assist smartwatch users in understanding their movement routines more quickly and accurately.

Adaptive AI for context and activity recognition remains a relatively unexplored field due to difficulty in collecting sufficient information to develop supervised models. Additionally, building a dataset for human context activities “in the wild” demands time and human resources, which explains the lack of public datasets available. Some of the available datasets for activity recognition were collected using wearable sensors, since they are less invasive than images and precisely capture a user’s movements in time series. However, frequency series contain more information about sensors’ signals. In this paper, we investigate the use of feature engineering to improve the performance of a Deep Learning model. Thus, we propose using Fast Fourier Transform algorithms to extract features from frequency series instead of time series. We evaluated our approach on the ExtraSensory and WISDM datasets. The results show that using Fast Fourier Transform algorithms to extract features performed better than using statistics measures to extract features from temporal series. Additionally, we examined the impact of individual sensors on identifying specific labels and proved that incorporating more sensors enhances the model’s effectiveness. On the ExtraSensory dataset, the use of frequency features outperformed that of time-domain features by 8.9 p.p., 0.2 p.p., 39.5 p.p., and 0.4 p.p. in Standing, Sitting, Lying Down, and Walking activities, respectively, and on the WISDM dataset, the model performance improved by 1.7 p.p., just by using feature engineering.

Object detection is a widely explored topic within the computer vision research field mostly because it is necessary for almost every system containing some kind of visual scene understanding or interpretation. Significant advances throughout the last 40 years allowed us to evolve from early techniques based on template matching to modern deep detectors capable of detecting thousands of different classes of objects with reasonable performance. Nonetheless, as approaches kept improving, more challenging topics related to object detection have been proposed. Classic object detectors have to be trained with all classes that might be presented in the testing phase. However, this is a problem in real-world scenarios because it is impossible to know the whole domain of possible objects. Hence, a task has emerged called class-agnostic object detection that essentially detects objects without determining their classes. In this paper, we address this task using a convolutional network and texture graylevel quantization. Our results showed that our model could improve 2.1 percentage points (p.p.) from the best baseline on objects that were not annotated in the training phase.

Point cloud compression has been studied in standard bodies and we are here concerned with the Moving Picture Experts Group video-based point cloud compression (V-PCC) solution. Plenoptic point clouds (PPC) is a novel volumetric data representation wherein points are associated with colors in all viewing directions to improve realism. It is sampled as a number ( Nc ) of attribute colors per point. We propose a new method for the efficient video-based compression of PPC that is backwards compatible with the existing single-color V-PCC decoder. V-PCC generates three image atlases which are encoded using an image/video encoder. We assume there may be a reference color which is to be encoded as the main payload. We generate Nc+3 atlases and we produce Nc differential images against the reference color image. Those difference images are pixel-wise transformed using an Nc -point discrete cosine transform, generating Nc transformed atlases which are encoded, forming the secondary payload. Such secondary information is the plenoptic enhancement to the point cloud. If there is no reference attribute, we skip the differences and use the lowest frequency of the transformed atlases as the main payload. Results are presented that show an unrivaled performance of the proposed method.

Generating novel views from a single input is a challenging task that requires the prediction of occluded and non-visible content. Nevertheless, it is an interesting and active area of research due to its several applications such as entertainment. In this work, we propose an end-to-end architecture for monocular view synthesis based on the layered scene inference (LSI) method. The LSI uses layered depth images that can represent complex scenes with a reduced number of layers. To improve the LSI predictions, we develop two new strategies: (i) a pyramidal architecture that learns LDI predictions for different resolutions of the input and (ii) an image outpainting for filling the missing information at the LDI borders. We evaluate our method on the KITTI dataset, and show that the proposed versions outperform the baseline.

Anomaly-based approaches for detecting network intrusions suffer from accurate evaluation, comparison, and deployment due to the scarcity of adequate datasets. Consequently, researchers resort to suboptimal datasets that no longer relate to a real-world network nor provide insights for current network issues, such as the DARPA'98 dataset and its variants KDD'99 and NSL-KDD. In this article, we propose a statistical study over the NSL-KDD features, and we conclude that NSL-KDD and the old KDD'99 should not be used as a benchmark for creating novel anomaly-based approaches intrusion detection systems because they introduce a biased classification, since features are over-correlated. The proposed approach analyzes the correlation among features instead of checking for redundant values or the imbalance of data. Our results align with the performance of three machine learning techniques trained to discriminate attack from normal traffic. We show that biased classification occurs because there was a high correlation between features and classes. The syntactically-generated features are statistically different between normal and attack traffic classes, which implies that, in KDD-related datasets, attackers are not stealthy.