Submit to Special Issue Submit Abstract to Special Issue Review for Applied Sciences Propose a Special Issue

Journal Menu

Journal Browser

Machine Perception and Learning

Print Special Issue Flyer
Special Issue Editors
Special Issue Information
Keywords
Benefits of Publishing in a Special Issue
Published Papers

A special issue of Applied Sciences (ISSN 2076-3417). This special issue belongs to the section "Computing and Artificial Intelligence".

Deadline for manuscript submissions: 20 July 2025 | Viewed by 8010

Share This Special Issue

Special Issue Editor

Prof. Dr. Yi Ding

E-Mail Website
Guest Editor

School of Information and Software Engineering, University of Electronic Science and Technology of China, Chengdu 610054, China
Interests: machine learning; image analysis

Special Issue Information

Dear Colleagues,

Machine perception and learning are highly interdisciplinary and draw on findings in psychology, neuroscience, machine learning, computer vision, and behavioral economics. The mission of this field is to enable machines to perceive and understand the real world in order for them to intelligently generate multimodal content and perform robustly in challenging tasks. Recently, researchers have started to apply a range of machine learning- and AI-based methods to a wide variety of data sources, including multispectral, medical imagery, camera images, live webcam streams and video data. The recurring objective is to design efficient and accurate algorithms for the automatic extraction of semantic information from the data source. There is clear scope for the further development of such approaches to enhance the performance of associated technologies, which is the key aim of this journal, such as machine learning, deep learning, and transfer learning methods and AI models.

We welcome original and well-grounded research papers on all aspects of the foundations of machine perception and learning. The contributions may be theoretical, methodological, algorithmic, empirical, integrative (connecting ideas and methods across machine perception and learning), or critical (e.g., principled analyses and arguments that draw attention to goals, assumptions, or approaches). The submissions should place emphasis on the demonstrated or potential impact of the research in addressing pressing societal challenges, e.g., health, food, environment, education, governance, among others. All submissions will be evaluated and scored for the significance and novelty of the contributions (research problems or questions addressed, methods, experiments, analyses), theoretical and/or empirical soundness of the claims, and clarity of exposition.

The topics of interest include, but are not limited to:

AI-related brain and cognitive science;
Machine perception and human–machine interaction;
Machine learning and data mining;
Multimodal emotion recognition;
Pattern recognition and computer vision;
Signal processing and recognition;
Medical image processing;
Semi-supervised and weakly supervised learning;
Intelligent information processing;
Natural language processing;
Network intelligence and mobile computing;
Intelligent control and decision;
Robotics and intelligent systems;
Auto-ML;
Information fusion from disparate sources.

Prof. Dr. Yi Ding
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Applied Sciences is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

AI-related brain and cognitive science
machine perception and human–machine interaction
machine learning and data mining
multimodal emotion recognition
pattern recognition and computer vision
signal processing and recognition
medical image processing
semi-supervised and weakly supervised learning
intelligent information processing
natural language processing
network intelligence and mobile computing
intelligent control and decision
robotics and intelligent systems
auto-ML
information fusion from disparate sources

Benefits of Publishing in a Special Issue

Ease of navigation: Grouping papers by topic helps scholars navigate broad scope journals more efficiently.
Greater discoverability: Special Issues support the reach and impact of scientific research. Articles in Special Issues are more discoverable and cited more frequently.
Expansion of research network: Special Issues facilitate connections among authors, fostering scientific collaborations.
External promotion: Articles in Special Issues are often promoted through the journal's social media, increasing their visibility.
e-Book format: Special Issues with more than 10 articles can be published as dedicated e-books, ensuring wide and rapid dissemination.

Further information on MDPI's Special Issue policies can be found here.

Published Papers (5 papers)

Download All Papers

Order results

Result details

Show export options Show export options

Select all

Export citation of selected articles as:

Research

17 pages, 3658 KiB

Open AccessArticle

Change and Detection of Emotions Expressed on People’s Faces in Photos

by Zbigniew Piotrowski, Maciej Kaczyński and Tomasz Walczyna

Appl. Sci. 2024, 14(22), 10681; https://doi.org/10.3390/app142210681 - 19 Nov 2024

Viewed by 960

Abstract

Human emotions are an element of attention in various areas of interest such as psychology, marketing, medicine, and public safety. Correctly detecting human emotions is a complex matter. The more complex and visually similar emotions are, the more difficult they become to distinguish. Making visual modifications to the faces of people in photos in a way that changes the perceived emotion while preserving the characteristic features of the original face is one of the areas of research in deepfake technologies. The aim of this article is to showcase the outcomes of computer simulation experiments that utilize artificial intelligence algorithms to change the emotions on people’s faces. In order to detect and change emotions, deep neural networks discussed further in this article were used. Full article

(This article belongs to the Special Issue Machine Perception and Learning)

► Show Figures

Figure 1

Figure 1
Wheel of emotion [<a href="#B3-applsci-14-10681" class="html-bibr">3</a>]. Full article ">Figure 2
Circumplex theory of affect [<a href="#B3-applsci-14-10681" class="html-bibr">3</a>]. Full article ">Figure 3
EmoDNN emotion change preview. Full article ">Figure 4
Confusion matrices of trained classifiers (from left based on PyTorch; from right based on TensorFlow). Full article ">Figure 5
Confusion matrices of trained classifiers of generated faces with changed emotion (from left based on PyTorch; from right based on TensorFlow). Full article ">Figure A1
Preview of sample generated images for individual emotions (viewed from the top, the rows represent different emotions; viewed from the left, the consecutive columns represent pairs of images: [original image, image with changed emotion generated by EmoDNN]). Full article ">Figure A1 Cont.
Preview of sample generated images for individual emotions (viewed from the top, the rows represent different emotions; viewed from the left, the consecutive columns represent pairs of images: [original image, image with changed emotion generated by EmoDNN]). Full article ">

14 pages, 3539 KiB

Open AccessArticle

Discrimination Ability and Concentration Measurement Accuracy of Effective Components in Aroma Essential Oils Using Gas Sensor Arrays with Machine Learning

by Toshio Itoh, Pil Gyu Choi, Yoshitake Masuda, Woosuck Shin, Junichirou Arai and Nobuaki Takeda

Appl. Sci. 2024, 14(19), 8859; https://doi.org/10.3390/app14198859 - 2 Oct 2024

Viewed by 990

Abstract

Aroma essential oils contain ingredients that are beneficial to the human body. A gas sensor array is required to monitor the concentration of these essential oil components to regulate their concentration by air conditioning systems. Therefore, we investigated the discrimination ability and concentration measurement accuracy of 14 effective components, including four aroma essential oils (lavender, melissa, tea tree, and eucalyptus), from a single gas sample and mixtures of two gases using sensor arrays. To obtain our data, we used two sensor arrays comprising commercially available semiconductor sensors and our developed semiconductor sensors. For machine learning, principal component analysis was used to visualize the dataset obtained from the sensor signals, and an artificial neural network was used for a detailed analysis. Our developed sensor array, which included sensors that possessed excellent sensor responses to 14 effective components and combined different semiconductive sensor principles, showed a better discrimination and prediction accuracy than the commercially available sensors investigated in this study. Full article

(This article belongs to the Special Issue Machine Perception and Learning)

► Show Figures

Figure 1

Figure 1
Structural formulae of the effective components as target gases: (1) terpinen-4-ol, (2) α-terpinene, (3) γ-terpinene, (4) α-terpineol, (5) eucalyptol, (6) d-limonene, (7) α-pinene, (8) β-pinene, (9) p-cymene, (10) linalool, (11) linalyl acetate, (12) citronellal, and (13) citral (mixture of cis [neral]-trans [geranial] isomers). Full article ">Figure 2
Model showing the responses of eight sensors to the target gas and the location of the data used for data analysis. Yellowish region indicates target gas flowing into the sensor chamber. Full article ">Figure 3
Model of ANN used in the study. Full article ">Figure 4
Dynamic resistance responses: sensor array and target gas are (a) C and single gas No. 12, (b) L and single gas No. 12, (c) C and single gas No. 6, (d) L and single gas No. 6, (e) C and double gases Nos. 2 + 6, and (f) L and double gases Nos. 2 + 6, respectively. Concentration levels of the target gases are, in order, (a–d) Lvs. 4, 3, 2, and 1; (e,f) Lvs. 3 + 3, 3 + 1, 1 + 1, and 1 + 3. Full article ">Figure 5
PCA scores and eigenvectors: sensor array and dataset are (a) C and 1 min data, (b) L and 1 min data, (c) C and 12 min data, and (d) L and 12 min data, respectively. Numbers on plots from Lv. 4 indicate gas numbers. Full article ">Figure 6
Relationship diagram between true and predicted concentrations for target gas No. 12 using (a,b) 1 min data and (c,d) 12 min data on sensor arrays (a,c) C and (b,d) L. Plot colors are black: single gas, blue: highest concentration component of double gases, green: second highest concentration component of double gas, and red: other double gases. Full article ">Figure 7
Relationship diagram between true and predicted concentrations for target gas No. 6 using (a,b) 1 min data and (c,d) 12 min data on sensor arrays (a,c) C and (b,d) L. Plot colors are black: single gas, blue: highest concentration component of double gases, green: second highest concentration component of double gas, and red: other double gases. Full article ">

20 pages, 3088 KiB

Open AccessArticle

Passive TDOA Emitter Localization Using Fast Hyperbolic Hough Transform

by Gyula Simon and Ferenc Leitold

Appl. Sci. 2023, 13(24), 13301; https://doi.org/10.3390/app132413301 - 16 Dec 2023

Cited by 3 | Viewed by 1426

Abstract

A fast Hough transform (HT)-based hyperbolic emitter localization system is proposed to process time difference of arrival (TDOA) measurements. The position-fixing problem is provided for cases where the source is known to be on a given plane (i.e., the elevation of the source is known), while the sensors can be deployed anywhere in the three-dimensional space. The proposed solution provides fast evaluation and guarantees the determination of the global optimum. Another favorable property of the proposed solution is that it is robust against faulty sensor measurements (outliers). A fast evaluation method involving the hyperbolic Hough transform is proposed, and the global convergence property of the algorithm is proven. The performance of the algorithm is compared to that of the least-squares solution, other HT-based solutions, and the theoretical limit (the Cramér–Rao lower bound), using simulations and real measurement examples. Full article

(This article belongs to the Special Issue Machine Perception and Learning)

► Show Figures

Figure 1

17 pages, 12887 KiB

Open AccessArticle

Deep Neural Network-Based Autonomous Voltage Control for Power Distribution Networks with DGs and EVs

by Durim Musiqi, Vjosë Kastrati, Alessandro Bosisio and Alberto Berizzi

Appl. Sci. 2023, 13(23), 12690; https://doi.org/10.3390/app132312690 - 27 Nov 2023

Cited by 6 | Viewed by 1740

Abstract

This paper makes use of machine learning as a tool for voltage regulation in distribution networks that contain electric vehicles and a large production from distributed generation. The methods of voltage regulation considered in this study are electronic on-load tap changers and line voltage regulators. The analyzed study-case represents a real-life feeder which operates at 10 kV. It has 9 photovoltaic systems with various peak installed powers, 2 electric vehicle charging stations, and 41 secondary substations, each with an equivalent load. Measurement data of loads and irradiation data of photovoltaic systems were collected hourly for two years. Those data are used as inputs in the feeder’s model in DigSilent PowerFactory where Quasi-Dynamic simulations are run. That will provide the correct tap positions as outputs. These inputs and outputs will then serve to train a Deep Neural Network which later will be used to predict the correct tap positions on input data it has not seen before. Results show that ML in general and DNN specifically show usefulness and robustness in predicting correct tap positions with very small computational requirements. Full article

(This article belongs to the Special Issue Machine Perception and Learning)

► Show Figures

Figure 1

18 pages, 4434 KiB

Open AccessArticle

Enhancing Anomaly Detection Models for Industrial Applications through SVM-Based False Positive Classification

by Ji Qiu, Hongmei Shi, Yuhen Hu and Zujun Yu

Appl. Sci. 2023, 13(23), 12655; https://doi.org/10.3390/app132312655 - 24 Nov 2023

Cited by 5 | Viewed by 2208

Abstract

Unsupervised anomaly detection models are crucial for the efficiency of industrial applications. However, frequent false alarms hinder the widespread adoption of unsupervised anomaly detection, especially in fault detection tasks. To this end, our research delves into the dependence of false alarms on the baseline anomaly detector by analyzing the high-response regions in anomaly maps. We introduce an SVM-based false positive classifier as a post-processing module, which identifies false alarms from positive predictions at the object level. Moreover, we devise a sample synthesis strategy that generates synthetic false positives from the trained baseline detector while producing synthetic defect patch features from fuzzy domain knowledge. Following comprehensive evaluations, we showcase substantial performance enhancements in two advanced out-of-distribution anomaly detection models, Cflow and Fastflow, across image and pixel-level anomaly detection performance metrics. Substantive improvements are observed in two distinct industrial applications, with notable instances of elevating the image-level F1-score from 46.15% to 78.26% in optimal scenarios and boosting pixel-level AUROC from 72.36% to 94.74%. Full article

(This article belongs to the Special Issue Machine Perception and Learning)

► Show Figures

Figure 1

Figure 1
Detection results of an OOD model without and with the proposed false-positive classifier on a wood defect detection task. The baseline segmentation model is Fastflow [<a href="#B14-applsci-13-12655" class="html-bibr">14</a>], consisting of a deep feature extraction backbone initialized with the ImageNet pre-trained weights and a normalizing flow network trained by the anomaly-free wood images. Parameters of baseline segmentation models freeze during the testing process. Full article ">Figure 2
Density probability distributions of prediction scores. (a) depicts the distribution of anomaly-free targets <math display="inline"><semantics> <mrow> <msub> <mrow> <mi>P</mi> </mrow> <mrow> <mi>K</mi> <mn>1</mn> </mrow> </msub> </mrow> </semantics></math> produced by the segmentation model in the training process. (b) shows the ideal condition of OOD models: When the distribution of the anomalies <math display="inline"><semantics> <mrow> <msub> <mrow> <mi>P</mi> </mrow> <mrow> <mi>K</mi> <mn>2</mn> </mrow> </msub> </mrow> </semantics></math> is shown as a solid orange line, one threshold exists for successful detection with a perfect AUROC score. Similarly, more thresholds exist for distant distributions like the dotted orange line. (c) presents an actual condition that distributions intersect due to inferior discrimination ability. False alarms arise and are present in the red region. Full article ">Figure 3
The proposed optimization workflow. Blue arrows describe the baseline defect detection processes that directly generate outputs from the baseline segmentation model. Orange arrows present workflows of the devised post-processing method, which filters out false alarms from candidate positive patches. Green arrows draw the sample synthesis and model training process of the SVM classification model depicted in <a href="#sec3dot3-applsci-13-12655" class="html-sec">Section 3.3</a>. Full article ">Figure 4
Sample synthesis workflow. Training samples for the classifier are represented as vectors, with their dimensions tailored to the selected discriminative prior knowledge descriptions. Synthetic defect samples are generated based on fuzzy knowledge, while synthetic false alarm samples are derived from high-response regions within the anomaly-free training dataset. Full article ">Figure 5
Anomaly maps of anomaly-free images within a training dataset. Full article ">Figure 6
Visualization of comparative experiments on wood defect examination using Fastflow. Columns correspond to: (a) the test image, (b) ground truth, (c) anomaly map from the baseline OOD model, (d) baseline defect mask, (e) baseline segmentation result, (f) filtered mask, and (g) filtered segmentation result. Full article ">Figure 7
Visualization of comparison experiments on wood defect examination using Cflow. Columns correspond to: (a) the test image, (b) ground truth, (c) anomaly map from the baseline OOD model, (d) baseline defect mask, (e) baseline segmentation result, (f) filtered mask, and (g) filtered segmentation result. Full article ">Figure 8
Visualization of comparison experiments on the round pin examination of a freight train using Fastflow. Columns correspond to: (a) the test image, (b) ground truth, (c) anomaly map from the baseline OOD model, (d) baseline defect mask, (e) baseline segmentation result, (f) filtered mask, and (g) filtered segmentation result. Full article ">Figure 9
Visualization of comparison experiments on a round pin examination of a freight train using Cflow. Columns correspond to: (a) the test image, (b) ground truth, (c) anomaly map from the baseline OOD model, (d) baseline defect mask, (e) baseline segmentation result, (f) filtered mask, and (g) filtered segmentation result. Full article ">

Show export options Show export options

Select all

Export citation of selected articles as:

Displaying articles 1-5

Journal Menu

Journal Browser

Machine Perception and Learning

Share This Special Issue

Special Issue Editor

Special Issue Information

Keywords

Benefits of Publishing in a Special Issue

Published Papers (5 papers)

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI