Towards an efficient and robust foot classification from pedobarographic images

This study aimed to evaluate the agreement between radiographic and pedobarographic measurement methods for classifying a normal foot type following calcaneal lengthening osteotomy in children with symptomatic pes planus. Thirty feet from patients (mean age: 15.1 years; range: 11-17 years) with idiopathic symptomatic hypermobile pes planus who had data and over 1 year of follow-up were evaluated. Radiographic assessments were categorized into measurements evaluating the medial longitudinal arch and those assessing calcaneal morphology. Medial longitudinal arch measurements included talocalcaneal, lateral talo-first metatarsal, talohorizontal, calcaneal pitch, talonavicular coverage, anteroposterior talo-first metatarsal, and Moreau-Costa-Bartani angles, while calcaneal measurements included calcaneal length, posterior facet height, Gissane angle, and Böhler angle. Pedobarographic measurements were performed using Arch Index, Footprint Index, and Clarke's angle, and clinical outcomes were assessed using the Oxford Ankle Foot Questionnaire for Children. Cohen's kappa coefficient was used to determine the agreement among the measurement methods in classifying the foot type as 'normal'. Postoperative findings revealed that changes in calcaneal length occurred without deformation of the articular surface. Arch Index exhibited 'almost perfect' agreement with the lateral talo-first metatarsal (k = 0.902) and talohorizontal angles (k = 0.814) and was the only pedobarographic measurement that correlated significantly with all domains of the Oxford Ankle Foot Questionnaire for Children. In contrast, Clarke's angle and Footprint Index did not show agreement with radiographic measurements. Following calcaneal lengthening osteotomy, among the pedobarographic measurement methods, only the Arch Index can be used as an alternative to radiographic measurement methods for classifying normal foot structure.

The human foot has frequently been categorized into arch height groups based upon analysis of footprint parameters. This study investigates the relationship between directly measured arch height and many of the footprint parameters that have been assumed to represent arch height. A total of 115 male subjects were measured and footprint parameters were calculated from digitized outlines. Correlation and regression analyses were used to determine the relationship between footprint measures and arch height. It may be concluded from the results that footprint parameters proposed in the literature (arch angle, footprint index, and arch index) and two further parameters suggested in this study (arch length index and truncated arch index) are invalid as a basis for prediction or categorization of arch height. The categorization of the human foot according to the footprint measures evaluated in this paper represent no more than indices and angles of the plantar surface of the foot itself.

Attempts to evaluate foot arch types from footprint parameters have yielded conflicting results in the past. This could be caused by the uncertainty inherent in the definition of some footprint parameters and the inaccuracy during the footprint acquisition and the parameter calculation phases of the traditional methods. In order to avoid these problems, digital image processing methods were used to acquire and to calculate the Arch Index (AI), a parameter which is robust in its definition. A significant correlation (r = -0.70, p < 0.0001) was found between AI and arch height. Therefore this study confirms that foot arch type does correlate with the footprint parameter, AI. This was further revealed by a new parameter, the Modified Arch Index (MAI), which incorporates foot pressure information in the evaluation. MAI not only correlated well with arch height (r = -0.71, p < 0.0001) but appeared to characterize abnormal foot types better than AI.

The Reliability of the Associate Platinum Digital Foot Scanner in Measuring Previously Developed Footprint Characteristics: A Technical Note

Ultrasonic cleaning equipment

Accurate sleep stage scoring is crucial for diagnosing and treating sleep disorders, yet traditional manual methods are time-consuming and susceptible to variability. While recent advancements in machine learning and deep learning have enhanced automated sleep stage detection, many approaches still rely on handcrafted features and encounter limitations when processing full-night data. In this paper, we introduce a novel many-to-many classification framework that leverages a temporal focal modulation network for efficient and accurate sleep stage scoring. Our model, SleepFocalNet, processes full-night single-channel EEG signals and predicts sleep stages for all epochs simultaneously. SleepFocalNet is composed of three key components: a convolution block for local feature extraction, a focal modulation block for long-range temporal modeling, and a classification block for final predictions. We evaluated SleepFocalNet on Sleep Heart Health Study (SHHS), SleepEDF-20, and SleepEDF-78 datasets, achieving state-of-the-art performance. On SHHS, SleepFocalNet attained an accuracy of 0.888 and an F1-score of 0.815. On SleepEDF-20, it obtained an accuracy of 0.885 and an F1-score of 0.836. On SleepEDF-78, it outperformed other models with an accuracy of 0.855 and an F1-score of 0.800. This study represents the first application of temporal focal modulation networks in sleep stage scoring. Additionally, we conducted an extensive analysis of various network configurations to assess the impact of different architectural choices on performance. The results validate the potential of our approach to enhance the reliability and scalability of automated sleep stage scoring, offering a robust alternative to existing methods.

Sea ice plays a critical role in the global climate system and maritime operations, making timely and accurate classification essential. However, traditional manual methods are time-consuming, costly, and have inherent biases. Automating sea-ice type classification addresses these challenges by enabling faster, more consistent, and scalable analysis. While both traditional and deep-learning approaches have been explored, deep-learning models offer a promising direction for improving efficiency and consistency in sea-ice classification. However, the absence of a standardized benchmark and comparative study prevents a clear consensus on the best-performing models. To bridge this gap, we introduce IceBench, a comprehensive benchmarking framework for sea-ice type classification. Our key contributions are three-fold: First, we establish the IceBench benchmarking framework, which leverages the existing AI4Arctic Sea Ice Challenge Dataset as a standardized dataset, incorporates a comprehensive set of evaluation metrics, and includes representative models from the entire spectrum of sea-ice type-classification methods categorized in two distinct groups, namely pixel-based classification methods and patch-based classification methods. IceBench is open-source and allows for convenient integration and evaluation of other sea-ice type-classification methods, hence facilitating comparative evaluation of new methods and improving reproducibility in the field. Second, we conduct an in-depth comparative study on representative models to assess their strengths and limitations, providing insights for both practitioners and researchers. Third, we leverage IceBench for systematic experiments addressing key research questions on model transferability across seasons (time) and locations (space), data downsampling, and preprocessing strategies. By identifying the best-performing models under different conditions, IceBench serves as a valuable reference for future research and a robust benchmarking framework for the field.

We focus the attention on the area of generic and automatic audio classification and segmentation for audio-based multimedia indexing and retrieval applications. In particular, we present a fuzzy approach toward hierarchic audio classification and global segmentation framework based on automatic audio analysis providing robust, bi-modal, efficient and parameter invariant classification over global audio segments. The input audio is split into segments, which are classified as speech, music, fuzzy or silent. The proposed method minimizes critical errors of misclassification by fuzzy region modeling, thus increasing the efficiency of both pure and fuzzy classification. The experimental results show that the critical errors are minimized and the proposed framework significantly increases the efficiency and the accuracy of audio-based retrieval especially in large multimedia databases.

Sound-based tinnitus intervention stimuli include broad-band noise signals with subjectively adjusted bandwidths used as maskers delivered by commercial devices or hearing aids, environmental sounds broadly described and delivered by both consumer devices and hearing aids, music recordings specifically modified and delivered in a variety of different ways, and other stimuli. Acoustic coordinated reset neuromodulation therapy for tinnitus reduction has unique and more stringent requirements compared to all other sound-based tinnitus interventions. These include precise characterization of tinnitus pitch and loudness, and effective provision of patient-controlled daily therapy signals at defined frequencies, levels, and durations outside of the clinic. The purpose of this study was to evaluate an approach to accommodate these requirements including evaluation of a mobile device, validation of an automated tinnitus pitch-matching algorithm and assessment of a patient's ability to control stimuli and collect repeated outcome measures. The experimental design involved direct laboratory measurements of the sound delivery capabilities of a mobile device, comparison of an automated, adaptive pitch-matching method to a traditional manual method and measures of a patient's ability to understand and manipulate a mobile device graphic user interface to both deliver the therapy signals and collect the outcome measures. This study consisted of 5 samples of a common mobile device for the laboratory measures and a total of 30 adult participants: 15 randomly selected normal-hearing participants with simulated tinnitus for validation of a tinnitus pitch-matching algorithm and 15 sequentially selected patients already undergoing tinnitus therapy for evaluation of patient usability. No tinnitus intervention(s) were specifically studied as a component of this study. Data collection involved laboratory measures of mobile devices, comparison of manual and automated adaptive tinnitus pitch-matching psychoacoustic procedures in the same participant analyzed for absolute differences (t test), variance differences (f test), and range comparisons, and assessment of patient usability including questionnaire measures and logs of patient observations. Mobile devices are able to reliably and accurately deliver the acoustic therapy signals. There was no difference in mean pitch matches (t test, p > 0.05) between an automated adaptive method compared to a traditional manual pitch-matching method. However, the variability of the automated pitch-matching method was much less (f test, p < 0.05) with twice as many matches within the predefined error range (±5%) compared to the manual pitch-matching method (80% versus 40%). After a short initial training, all participants were able to use the mobile device effectively and to perform the required tasks without further professional assistance.

INTRODUCTION: Although baropodometric analysis has been published since the 1990s, only now it is found a considerable number of studies showing different uses in the rehabilitation. OBJECTIVE: To amplify the use of this technology, this research aimed to analyze baropodometric records during upright position of subjects with hemiparesis, describing a way to define weight-bearing profiles in this population. METHOD: 20 healthy subjects were matched by gender and age with 12 subjects with chronic spastic hemiparesis. This control group was formed to establish the limits of symmetry during weight-bearing distribution in the hemiparesis group. Next, hemiparesis group was submitted to procedures to measure baropodometric records used to provide variables related to the weight-bearing distribution, the arch index and the displacements in the center of pressure (CoP). Data were used to compare differences among kinds of weight-bearing distribution (symmetric, asymmetric toward non-paretic or paretic foot) and coordination system for CoP displacements. RESULTS: Hemiparesis group was compounded by eight symmetrics, eight asymmetrics toward non-paretic foot and four asymmetric toward paretic foot. Significant differences in the weight-bearing distributions between non-predominantly and predominantly used foot did not promote differences in the other baropodometric records (peak and mean of pressure, and support area). Mainly in the asymmetry toward non-paretic foot it was observed significant modifications of the baropodometric records. CONCLUSION: Baropodometric technology can be used to analyze weight-bearing distribution during upright position of subjects with hemiparesis, detecting different kinds of weight-bearing profiles useful to therapeutic programs and researches involving subjects with this disability.

This study was carried out to assess whether the spatial resolution has an impact on the detection accuracy of proximal caries in flat panel CBCT (cone beam computerized tomography) images and if the detection accuracy can be improved by flat panel CBCT images scanned with high spatial resolution when compared to digital intraoral images. The CBCT test images of 45 non-restored human permanent teeth were respectively scanned with the ProMax 3D and the DCT Pro scanners at different resolutions. Digital images were obtained with a phosphor plate imaging system Digora Optime. Eight observers evaluated all the test images for carious lesion within the 90 proximal surfaces. With the histological examination serving as the reference standard, observer performances were evaluated by receiver operating characteristic (ROC) curves. The areas under the ROC curves were analyzed with two-way analysis of variance. No significant differences were found among the CBCT images and between CBCT and digital images when only proximal enamel caries was detected (p = 0.989). With respect to the detection of proximal dentinal caries, significant difference was found between CBCT and digital images (p < 0.001) but not among CBCT images. The spatial resolution did not have an impact on the detection accuracy of proximal caries in flat panel CBCT images. The flat panel CBCT images scanned with high spatial resolution did not improve the detection accuracy of proximal enamel caries compared to digital intraoral images. CBCT images scanned with high spatial resolutions could not be used for proximal caries detection.

Digital images have become an important component of cytology practice. They are used in telecytology, automated screening, educational material, and Web sites and have potential for use in proficiency testing. However, there has been no formal evaluation to date to determine if digital image manipulation (intentional or unintentional) can affect their interpretation. To investigate whether alteration of digital cytology images affects diagnosis. Acquired digital images of ThinPrep Papanicolaou test slides were manipulated (rotated 90 degrees and brightness, contrast, red-green-blue color, and luminosity adjusted) using Photoshop. A test composed of these altered images, along with their original (unaltered) image and exact duplicates was given to 22 cytologists (13 cytotechnologists, 8 cytopathologists, and 1 fellow). All images were rated as negative, atypical (atypical squamous cells of undetermined significance), low-grade squamous intraepithelial lesion, high-grade squamous intraepithelial lesion, or positive for cancer. Weighted kappa and heterogeneity chi(2) statistics were used to measure levels of agreement and assess concordance between groups. The level of agreement for identical duplicate images was excellent (kappa = 0.81), compared with the poor agreement for manipulated image pairs (kappa = 0.21), a statistically significant difference (P < .001). For all altered image types agreement was poor. There was no significant difference between cytotechnologists and cytopathologists in level of agreement (P = .56). Manipulation of a Papanicolaou test digital image, irrespective of the specific category of cytologic material photographed, significantly affects its interpretation by both cytotechnologists and cytopathologists. This suggests that care needs to be taken when digital cytology images are used, to specifically ensure that their alteration does not affect diagnosis.

BackgroundMore than 1000 scientific papers have been devoted to flatfoot issue. However, a bimodal distribution of flatfoot indices in school-aged children has never been discovered. The purposes of this study were to establish a new classification of flatfoot by characteristic in frequency distribution of footprint index and to endue the classification with discrepancy in physical fitness.Methods/Principal FindingsIn a longitudinal survey of physical fitness and body structure, weight bearing footprints and 3 physical fitness related tests were measured in 1228 school-aged children. Frequency distribution of initial data was tested by Kolmogorov-Smirnov test for normality and a unique bimodal distribution of footprint index was identified. The frequency distribution of footprint index manifests two distinct modes, flatfoot and non-flatfoot, by deconvolution and bootstrapping procedures. A constant intersection value of 1.0 in Staheli's arch index and 0.6 in Chippaux-Smirak index could distinguish the two modes of children, and the value was constant in different age, sex, and weight status. The performance of the one leg balance was inferior in flatfoot girls (median, 4.0 seconds in flatfoot girls vs. 4.3 seconds in non-flatfoot girls, p = 0.04, 95% CI 0.404–0.484).DiscussionThe natural bimodality lends itself to a flatfoot classification. Bimodality suggests development of the child's foot arch would be a leap from one state to another, rather than a continuous growth as body height and weight. The underlying dynamics of the human foot arch and motor development will trigger research prospects.

The automatic sound event classification (SEC) has attracted a growing attention in recent years. Feature extraction is a critical factor in SEC system, and the deep neural network (DNN) algorithms have achieved the state-of-the-art performance for SEC. The extreme learning machine-based auto-encoder (ELM-AE) is a new deep learning algorithm, which has both an excellent representation performance and very fast training procedure. However, ELM-AE suffers from the problem of unstability. In this work, a bilinear multi-column ELM-AE (B-MC-ELM-AE) algorithm is proposed to improve the robustness, stability, and feature representation of the original ELM-AE, which is then applied to learn feature representation of sound signals. Moreover, a B-MC-ELM-AE and two-stage ensemble learning (TsEL)-based feature learning and classification framework is then developed to perform the robust and effective SEC. The experimental results on the Real World Computing Partnership Sound Scene Database show that the proposed SEC framework outperforms the state-of-the-art DNN algorithm.

Although the Arabic language is spoken by more than 300 million people and is one of the six official languages of the United Nations (UN), there has been less research done on Arabic text data (compared to English) in the realm of machine learning, especially in text classification. In the past decade, Arabic data such as news, tweets, etc. have begun to receive some attention. Although automatic text classification plays an important role in improving the browsability and accessibility of data, Electronic Theses and Dissertations (ETDs) have not received their fair share of attention, in spite of the huge number of benefits they provide to students, universities, and future generations of scholars. There are two main roadblocks to performing automatic subject classification on Arabic ETDs. The first is the unavailability of a public corpus of Arabic ETDs. The second is the linguistic complexity of the Arabic language; that complexity is particularly evident in academic documents such as ETDs. To address these roadblocks, this paper presents Otrouha, a framework for automatic subject classification of Arabic ETDs, which has two main goals. The first is building a Corpus of Arabic ETDs and their key metadata such as abstracts, keywords, and title to pave the way for more exploratory research on this valuable genre of data. The second is to provide a framework for automatic subject classification of Arabic ETDs through different classification models that use classical machine learning as well as deep learning techniques. The first goal is aided by searching the AskZad Digital Library, which is part of the Saudi Digital Library (SDL). AskZad provides other key metadata of Arabic ETDs, such as abstract, title, and keywords. The current search results consist of abstracts of Arabic ETDs. This raw data then undergoes a pre-processing phase that includes stop word removal using the Natural Language Tool Kit (NLTK), and word lemmatization using the Farasa API. To date, abstracts of 518 ETDs across 12 subjects have been collected. For the second goal, the preliminary results show that among the machine learning models, binary classification (one-vs.-all) performed better than multiclass classification. The maximum per subject accuracy is 95%, with an average accuracy of 68% across all subjects. It is noteworthy that the binary classification model performed better for some categories than others. For example, Applied Science and Technology shows 95% accuracy, while the category of Administration shows 36%. Deep learning models resulted in higher accuracy but lower F-measure; their overall performance is lower than machine learning models. This may be due to the small size of the dataset as well as the imbalance in the number of documents per category. Work to collect additional ETDs will be aided by collaborative contributions of data from additional sources.