Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

A novel approach for automatic annotation of human actions in 3D point clouds for flexible collaborative tasks with industrial robots.

In Frontiers in robotics and AI

Manual annotation for human action recognition with content semantics using 3D Point Cloud (3D-PC) in industrial environments consumes a lot of time and resources. This work aims to recognize, analyze, and model human actions to develop a framework for automatically extracting content semantics. Main Contributions of this work: 1. design a multi-layer structure of various DNN classifiers to detect and extract humans and dynamic objects using 3D-PC preciously, 2. empirical experiments with over 10 subjects for collecting datasets of human actions and activities in one industrial setting, 3. development of an intuitive GUI to verify human actions and its interaction activities with the environment, 4. design and implement a methodology for automatic sequence matching of human actions in 3D-PC. All these procedures are merged in the proposed framework and evaluated in one industrial Use-Case with flexible patch sizes. Comparing the new approach with standard methods has shown that the annotation process can be accelerated by 5.2 times through automation.

Krusche Sebastian, Al Naser Ibrahim, Bdiwi Mohamad, Ihlenfeldt Steffen

2023

data labeling, deep learning, human activity recognition, point cloud annotation, robotics

Pathology Pathology

Decision-making support system for diagnosis of oncopathologies by histological images.

In Journal of pathology informatics ; h5-index 23.0

The aim of the study is to increase the functional efficiency of machine learning decision support system (DSS) for the diagnosis of oncopathology on the basis of tissue morphology. The method of hierarchical information-extreme machine learning of diagnostic DSS is offered. The method is developed within the framework of the functional approach to modeling of natural intelligence cognitive processes at formation and acceptance of classification decisions. This approach, in contrast to neuronal structures, allows diagnostic DSS to adapt to arbitrary conditions of histological imaging and flexibility in retraining the system by expanding the recognition classes alphabet that characterize different structures of tissue morphology. In addition, the decisive rules built within the geometric approach are practically invariant to the multidimensionality of the diagnostic features space. The developed method allows to create information, algorithmic, and software of the automated workplace of the histologist for diagnosing oncopathologies of different genesis. The machine learning method is implemented on the example of diagnosing breast cancer.

Dovbysh Anatoliy, Shelehov Ihor, Romaniuk Anatolii, Moskalenko Roman, Savchenko Taras

2023

Breast cancer, Computer-aided detection, Hierarchical information‐extreme machine learning, Histological image, Information criterion, Machine learning

General General

Improving the brain image resolution of generalized q-sampling MRI revealed by a three-dimensional CNN-based method.

In Frontiers in neuroinformatics

BACKGROUND : Understanding neural connections facilitates the neuroscience and cognitive behavioral research. There are many nerve fiber intersections in the brain that need to be observed, and the size is between 30 and 50 nanometers. Improving image resolution has become an important issue for mapping the neural connections non-invasively. Generalized q-sampling imaging (GQI) was used to reveal the fiber geometry of straight and crossing. In this work, we attempted to achieve super-resolution with a deep learning method on diffusion weighted imaging (DWI).

MATERIALS AND METHODS : A three-dimensional super-resolution convolutional neural network (3D SRCNN) was utilized to achieve super-resolution on DWI. Then, generalized fractional anisotropy (GFA), normalized quantitative anisotropy (NQA), and the isotropic value of the orientation distribution function (ISO) mapping were reconstructed using GQI with super-resolution DWI. We also reconstructed the orientation distribution function (ODF) of brain fibers using GQI.

RESULTS : With the proposed super-resolution method, the reconstructed DWI was closer to the target image than the interpolation method. The peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) were also significantly improved. The diffusion index mapping reconstructed by GQI also had higher performance. The ventricles and white matter regions were much clearer.

CONCLUSION : This super-resolution method can assist in postprocessing low-resolution images. With SRCNN, high-resolution images can be effectively and accurately generated. The method can clearly reconstruct the intersection structure in the brain connectome and has the potential to accurately describe the fiber geometry on a subvoxel scale.

Shin Chun-Yuan, Chao Yi-Ping, Kuo Li-Wei, Chang Yi-Peng Eve, Weng Jun-Cheng

2023

diffusion MRI, generalized q-sampling imaging (GQI), peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), super-resolution convolutional neural network (SRCNN)

General General

A study on the clusterability of latent representations in image pipelines.

In Frontiers in neuroinformatics

Latent representations are a necessary component of cognitive artificial intelligence (AI) systems. Here, we investigate the performance of various sequential clustering algorithms on latent representations generated by autoencoder and convolutional neural network (CNN) models. We also introduce a new algorithm, called Collage, which brings views and concepts into sequential clustering to bridge the gap with cognitive AI. The algorithm is designed to reduce memory requirements, numbers of operations (which translate into hardware clock cycles) and thus improve energy, speed and area performance of an accelerator for running said algorithm. Results show that plain autoencoders produce latent representations which have large inter-cluster overlaps. CNNs are shown to solve this problem, however introduce their own problems in the context of generalized cognitive pipelines.

Wheeldon Adrian, Serb Alexander

2023

artificial intelligence, autoencoders, clustering, cognitive computing, convolutional neural networks, machine learning, symbolics

General General

Determination of factors affecting customer satisfaction towards "maynilad" water utility company: A structural equation modeling-deep learning neural network hybrid approach.

In Heliyon

The Maynilad Water Services Inc. (MWSI) is responsible for supplying water to the west zone of Metro Manila. The utility provides service to 17 cities and municipalities which frequently experience water interruptions and price hikes. This study aimed to identify the key factors affecting customer satisfaction toward MWSI by integrating the SERVQUAL dimensions and Expectation Confirmation Theory (ECT). An online questionnaire was disseminated to 725 MWSI customers using the snowball sampling method to obtain accurate data. Ten latent were analyzed using Structural Equation Modeling and Deep Learning Neural Network hybrid. It was found that Assurance, Tangibles, Empathy, Expectations, Confirmation, Performance, and Water consumption were all factors affecting MWSI customers' satisfaction. Results showed that having an affordable water service, providing accurate water bills, on-time completion of repairs and installations, intermittent water interruptions and professional employees contribute to the general satisfaction. MWSI officials may utilize this study's findings to assess further the quality of their services and design effective policies to improve. The employment of DLNN and SEM hybrid showed promising results when employed in human behavior. Thus, the results of this study would be beneficial when examining satisfaction to utilities and policies among service providers in different countries. Moreover, this study could be extended and applied among other customer and service-focused industries worldwide.

Ong Ardvin Kester S, Prasetyo Yogi Tri, Sacro Mariela Celine C, Artes Alycia L, Canonoy Mariella Phoemela M, Onda Guia Karyl D, Persada Satria Fadil, Nadlifatin Reny, Robas Kirstien Paola E

2023-Mar

Deep learning neural network, Expectation-confirmation theory, Servqual, Structural equation modeling, Water utility

General General

A novel IEF-DLNN and multi-objective based optimizing control strategy for seawater reverse osmosis desalination plant.

In Heliyon

Over the past years, Seawater Desalination (SWD) has been enhanced regularly. In this desalination process, numerous technologies are available. The Reverse Osmosis (RO) process, which requires effectual control strategies, is the most commercially-dominant technology. Therefore, for SWD, a novel Interpolation and Exponential Function-centered Deep Learning Neural Network (IEF-DLNN) and multi-objective-based optimizing control system has been proposed in this research methodology. Initially, the input data are gathered; then, to control the desalination process, an optimal control technique has been utilized by employing Probability-centric Dove Swarm Optimization-Proportional Integral Derivative (PDSO-PID). The attributes of permeate are extracted before entering the RO process; after that, by utilizing the IEF-DLNN, the trajectory is predicted. For optimal selection, the extracted attributes are deemed if the trajectory is present, or else to mitigate energy consumption along with cost, the RO Desalination (ROD) is performed. In an experimental evaluation, regarding certain performance metrics, the proposed model's performance is analogized with the prevailing methodologies. The outcomes demonstrated that the proposed system achieved better performance.

Alghamdi Ahmed

2023-Mar

Fitness, Interpolation and exponential based deep learning neural network (IEF-DLNN), Iteration, Neural Network (NN), Probability-based dove swarm optimization-proportional integral derivative (PDSO-PID)