Compact Feature Research Articles

Enhancing Sound Source Localization via False Negative Elimination.

Sound source localization aims to localize objects emitting the sound in visual scenes. Recent works obtaining impressive results typically rely on contrastive learning. However, the common practice of randomly sampling negatives in prior arts can lead to the false negative issue, where the sounds semantically similar to visual instance are sampled as negatives and incorrectly pushed away from the visual anchor/query. As a result, this misalignment of audio and visual features could yield inferior performance. To address this issue, we propose a novel audio-visual learning framework which is instantiated with two individual learning schemes: self-supervised predictive learning (SSPL) and semantic-aware contrastive learning (SACL). SSPL explores image-audio positive pairs alone to discover semantically coherent similarities between audio and visual features, while a predictive coding module for feature alignment is introduced to facilitate the positive-only learning. In this regard SSPL acts as a negative-free method to eliminate false negatives. By contrast, SACL is designed to compact visual features and remove false negatives, providing reliable visual anchor and audio negatives for contrast. Different from SSPL, SACL releases the potential of audio-visual contrastive learning, offering an effective alternative to achieve the same goal. Comprehensive experiments demonstrate the superiority of our approach over the state-of-the-arts. Furthermore, we highlight the versatility of the learned representation by extending the approach to audio-visual event classification and object detection tasks.

U-Match: Exploring Hierarchy-Aware Local Context for Two-View Correspondence Learning.

Rejecting outlier correspondences is one of the critical steps for successful feature-based two-view geometry estimation, and contingent heavily upon local context exploration. Recent advances focus on devising elaborate local context extractors whereas typically adopting explicit neighborhood relationship modeling at a specific scale, which is intrinsically flawed and inflexible, because 1) severe outliers often populated in putative correspondences and 2) the uncertainty in the distribution of inliers and outliers make the network incapable of capturing adequate and reliable local context from such neighborhoods, therefore resulting in the failure of pose estimation. This prospective study proposes a novel network called U-Match that has the flexibility to enable implicit local context awareness at multiple levels, naturally circumventing the aforementioned issues that plague most existing studies. Specifically, to aggregate multi-level local context implicitly, a hierarchy-aware graph representation module is designed to flexibly encode and decode hierarchical features. Moreover, considering that global context always works collaboratively with local context, an orthogonal local-and-global information fusion module is presented to integrate complementary local and global context in a redundancy-free manner, thus yielding compact feature representations to facilitate correspondence learning. Thorough experimentation across relative pose estimation, homography estimation, visual localization, and point cloud registration affirms U-Match's remarkable capabilities.

An adaptive session-incremental broad learning system for continuous motor imagery EEG classification.

Motor imagery electroencephalography (MI-EEG) is usually used as a driving signal in neuro-rehabilitation systems, and its feature space varies with the recovery progress. It is required to endow the recognition model with continuous learning and self-updating capability. Broad learning system (BLS) can be remodeled in an efficient incremental learning way. However, its architecture is intractable to change automatically to adapt to new incoming MI-EEG with time-varying and complex temporal-spatial characteristics. In this paper, an adaptive session-incremental BLS (ASiBLS) is proposed based on mutual information theory and BLS. For the initial session data, a compact temporal-spatial feature extractor (CTS) is designed to acquire the temporal-spatial features, which are input to a baseline BLS (bBLS). Furthermore, for new session data, a mutual information maximization constraint (MIMC) is introduced into the loss function of CTS to make the features' probability distribution sufficiently similar to that of the previous session, a new incremental BLS sequence (iBLS) is obtained by adding a small number of nodes to the previous model, and so on. Experiments are conducted based on the BCI Competition IV-2a dataset with two sessions and IV-2b dataset with five sessions,ASiBLS achieves average decoding accuracies of 79.89% and 87.04%, respectively. The kappa coefficient and forgetting rate are also used to evaluate the model performance. The results show that ASiBLS can adaptively generate an optimized and reduced model for each session successively, which has better plasticity in learning new knowledge and stability in retaining old knowledge as well.

SSFAN: A Compact and Efficient Spectral-Spatial Feature Extraction and Attention-Based Neural Network for Hyperspectral Image Classification

Hyperspectral image (HSI) classification is a crucial technique that assigns each pixel in an image to a specific land cover category by leveraging both spectral and spatial information. In recent years, HSI classification methods based on convolutional neural networks (CNNs) and Transformers have significantly improved performance due to their strong feature extraction capabilities. However, these improvements often come with increased model complexity, leading to higher computational costs. To address this, we propose a compact and efficient spectral-spatial feature extraction and attention-based neural network (SSFAN) for HSI classification. The SSFAN model consists of three core modules: the Parallel Spectral-Spatial Feature Extraction Block (PSSB), the Scan Block, and the Squeeze-and-Excitation MLP Block (SEMB). After preprocessing the HSI data, it is fed into the PSSB module, which contains two parallel streams, each comprising a 3D convolutional layer and a 2D convolutional layer. The 3D convolutional layer extracts spectral and spatial features from the input hyperspectral data, while the 2D convolutional layer further enhances the spatial feature representation. Next, the Scan Block module employs a layered scanning strategy to extract spatial information at different scales from the central pixel outward, enabling the model to capture both local and global spatial relationships. The SEMB module combines the Spectral-Spatial Recurrent Block (SSRB) and the MLP Block. The SSRB, with its adaptive weight assignment mechanism in the SToken Module, flexibly handles time steps and feature dimensions, performing deep spectral and spatial feature extraction through multiple state updates. Finally, the MLP Block processes the input features through a series of linear transformations, GELU activation functions, and Dropout layers, capturing complex patterns and relationships within the data, and concludes with an argmax layer for classification. Experimental results show that the proposed SSFAN model delivers superior classification performance, outperforming the second-best method by 1.72%, 5.19%, and 1.94% in OA, AA, and Kappa coefficient, respectively, on the Indian Pines dataset. Additionally, it requires less training and testing time compared to other state-of-the-art deep learning methods.

Consistency-driven feature scoring and regularization network for visible–infrared person re-identification

Recently, visible–infrared person re-identification (VI-ReID) has received considerable attention due to its practical importance. A number of methods extract multiple local features to enrich the diversity of feature representations. However, some local features often involve modality-relevant information, leading to deteriorated performance. Moreover, existing methods optimize the models by only considering the samples at each batch while ignoring the learned features at previous iterations. As a result, the features of the same person images drastically change at different training epochs, hindering the training stability. To alleviate the above issues, we propose a novel consistency-driven feature scoring and regularization network (CFSR-Net), which consists of a backbone network, a local feature learning block, a feature scoring block, and a global–local feature fusion block, for VI-ReID. On the one hand, we design a cross-modality consistency loss to highlight modality-irrelevant local features and suppress modality-relevant local features for each modality, facilitating the generation of a reliable compact local feature. On the other hand, we develop a feature consistency regularization strategy (including a momentum class contrastive loss and a momentum distillation loss) to impose consistency regularization on the learning of different levels of features by considering the learned features at historical epochs. This effectively enables smooth feature changes and thus improves the training stability. Extensive experiments on public VI-ReID datasets clearly show the effectiveness of our method against several state-of-the-art VI-ReID methods. Code will be released at https://github.com/cxtjl/CFSR-Net.

Exploring Feature Representation Learning for Semi-Supervised Medical Image Segmentation.

This article presents a simple yet effective two-stage framework for semi-supervised medical image segmentation. Unlike prior state-of-the-art semi-supervised segmentation methods that predominantly rely on pseudo supervision directly on predictions, such as consistency regularization and pseudo labeling, our key insight is to explore the feature representation learning with labeled and unlabeled (i.e., pseudo labeled) images to regularize a more compact and better-separated feature space, which paves the way for low-density decision boundary learning and therefore enhances the segmentation performance. A stage-adaptive contrastive learning method is proposed, containing a boundary-aware contrastive loss that takes advantage of the labeled images in the first stage, as well as a prototype-aware contrastive loss to optimize both labeled and pseudo labeled images in the second stage. To obtain more accurate prototype estimation, which plays a critical role in prototype-aware contrastive learning, we present an aleatoric uncertainty-aware method to generate higher quality pseudo labels. Aleatoric-uncertainty adaptive (AUA) adaptively regularizes prediction consistency by taking advantage of image ambiguity, which, given its significance, is underexplored by existing works. Our method achieves the best results on three public medical image segmentation benchmarks.

Compact speckle spectrometer using femtosecond laser-induced double-sided surface nanostructures.

The utilization of light scattering in disordered media has shown promise in the design of highly sensitive speckle spectrometers. Significant advances have been made in the research of all-fiber speckle spectrometers, and various planar scattering media have also garnered the attention of many researchers. In this study, we designed a compact speckle spectrometer employing a femtosecond laser to induce double-sided nanostructures on a quartz glass as a scattering medium. Once the transmission matrix is calibrated, the spectrum can be reconstructed over a bandwidth of 100 nm, achieving a spectral resolution of 0.1 nm. A 5 pm spectral resolution has been demonstrated by integrating a neural network to recognize speckle patterns in 100 pm bandwidths at 1500, 1550, and 1600 nm. By combining ResNet-50 and GRU, a simulated continuous spectrum spinning a bandwidth of 2 nm can be accurately reconstructed. This innovative, compact spectrometer features low cost, small size, simple preparation, and repeatability.

TRust Your GENerator (TRYGEN): Enhancing Out-of-Model Scope Detection

Recent research has drawn attention to the ambiguity surrounding the definition and learnability of Out-of-Distribution recognition. Although the original problem remains unsolved, the term “Out-of-Model Scope” detection offers a clearer perspective. The ability to detect Out-of-Model Scope inputs is particularly beneficial in safety-critical applications such as autonomous driving or medicine. By detecting Out-of-Model Scope situations, the system’s robustness is enhanced and it is prevented from operating in unknown and unsafe scenarios. In this paper, we propose a novel approach for Out-of-Model Scope detection that integrates three sources of information: (1) the original input, (2) its latent feature representation extracted by an encoder, and (3) a synthesized version of the input generated from its latent representation. We demonstrate the effectiveness of combining original and synthetically generated inputs to defend against adversarial attacks in the computer vision domain. Our method, TRust Your GENerator (TRYGEN), achieves results comparable to those of other state-of-the-art methods and allows any encoder to be integrated into our pipeline in a plug-and-train fashion. Through our experiments, we evaluate which combinations of the encoder’s features are most effective for discovering Out-of-Model Scope samples and highlight the importance of a compact feature space for training the generator.

Hierarchical Keypoints Feature Alignment for Domain Adaptive Pose Estimation

Unsupervised Domain Adaptation (UDA) is a popular research topic in computer vision. One of the most common strategies in this field is to reduce domain shifts through global feature alignment. Unfortunately, we observe its failure on adaptive human pose estimation. From our analysis, we find out two major reasons: the extreme imbalance between keypoints and non-keypoints regions and the high diversity of human skeleton structures. To address these problems, we propose a simple yet effective approach named PoseDA. Our idea is to let the alignment focus on the feature of keypoint regions and enforce the model to learn domain-invariant representations for keypoint prediction. The key component of our PoseDA is a Hierarchical Feature Selection of Keypoints Regions (HFS) module, which consists of Coarse Feature Selection of Keypoints Regions (CFS) and Reliable Feature Selection of Keypoints Regions (RFS). By using HFS, we obtain reliable and compact keypoints features, allowing our model to achieve more effective feature alignment. Under three UDA scenarios, i.e. JTA→MPII, SURREAL→LSP, SURREAL→Human3.6, our PoseDA establishes new state-of-the-art performance. In particular, our PoseDA outperforms the previous best UDA methods by over 7% w.r.t. PCKh on JTA→MPII.

Leveraging interpretable machine learning in intensive care

AbstractIn healthcare, especially within intensive care units (ICU), informed decision-making by medical professionals is crucial due to the complexity of medical data. Healthcare analytics seeks to support these decisions by generating accurate predictions through advanced machine learning (ML) models, such as boosted decision trees and random forests. While these models frequently exhibit accurate predictions across various medical tasks, they often lack interpretability. To address this challenge, researchers have developed interpretable ML models that balance accuracy and interpretability. In this study, we evaluate the performance gap between interpretable and black-box models in two healthcare prediction tasks, mortality and length-of-stay prediction in ICU settings. We focus specifically on the family of generalized additive models (GAMs) as powerful interpretable ML models. Our assessment uses the publicly available Medical Information Mart for Intensive Care dataset, and we analyze the models based on (i) predictive performance, (ii) the influence of compact feature sets (i.e., only few features) on predictive performance, and (iii) interpretability and consistency with medical knowledge. Our results show that interpretable models achieve competitive performance, with a minor decrease of 0.2–0.9 percentage points in area under the receiver operating characteristic relative to state-of-the-art black-box models, while preserving complete interpretability. This remains true even for parsimonious models that use only 2.2 % of patient features. Our study highlights the potential of interpretable models to improve decision-making in ICUs by providing medical professionals with easily understandable and verifiable predictions.

Pavement compactness estimation based on 3D pavement texture features

Understanding the correlation between pavement compactness and pavement texture features aids in determining the range of compaction passes. This paper proposes an estimation model for pavement compactness based on 3D pavement features. The compaction process was simulated in laboratory for four different gradation types with varying compaction passes while 3D texture data were obtained. Parameters were calculated to interpret texture features. The Random Forest model and the Shapley additive explanations approach were used to explore the contribution of different feature parameters to the compactness prediction model for the feature selection. A polynomial linear model was proposed to predict the compactness using five selected parameters, which showed a good fit. It was also observed that Da and Spk make a more substantial contribution to the model. Additionally, an optimal compaction pass range considering compactness, texture depth, and temperature drop was proposed to support the control strategies in road compaction construction sites.

Estimating the all-terminal signatures for networks by using deep neural network

Computing the signature of a network is both significant and challenging. Addressing the limitations of existing methods in batch processing of large-scale network signatures, in this paper we propose a novel DNN (Deep Neural Network)-based framework for estimating the all-terminal signature and reliability for networks with varying topologies. Our framework involves constructing a DNN model with four efficient and compact network topological features (the numbers of nodes, and links, the node degrees and the link connectivity) as input features and the signature as the response. Additionally, we propose to estimate the all-terminal network reliability based on the signature estimated by the DNN, termed the two-stage DNN approach, which does not require the link reliability as one of the inputs, resulting in better estimation accuracy and generation performance compared to traditional DNN approaches. A case study is conducted and the results show that the estimation accuracy of our DNN model for the signature is satisfactory, and the two-stage DNN approach for network reliability outperforms existing DNN approaches in the literature.

Dissecting the planetary nebula NGC 4361 with MUSE

Optical integral field spectroscopy of planetary nebulae (PNe) offers a unique tool to explore the spatial relationships between the complex mixture of the many components (neutral, low- and high-ionisation gas, dust, and the central star) and their underlying physical conditions. The optical line and continuum emission in the very-high-ionisation Galactic PN, NGC 4361, were mapped to study the distribution of ionisation, extinction, electron temperature, and density. Based on commissioning data, MUSE Wide Field (60times 60$''$) normal-mode (4750-9300\ observations of NGC 4361 were reduced. The PN is larger than a single MUSE field and only the central 1 arcmin2 of the PN was observed in good conditions. Emission images in recombination and collisionally excited lines were extracted and the line ratios provided the dust extinction, electron density and temperature, and ionic abundances using standard techniques. A family of compact low-ionisation knots (dubbed 'freckles') was discovered and techniques developed to measure their spectra, independently of the extended high-ionisation medium. The nebula is confirmed as optically thin in the H-ionising continuum, based on its very low He i emission, even to the edges of the field. The electron temperature, $T_ e $, is shown to have a large-scale spatially coherent structure, as indicated by a previous long-slit spectrum. Prior to this study, no low-ionisation emission had been positively detected, although MUSE revealed both weak extended N ii and O ii and $>$100 spatially unresolved knots. There are several linear associations of these knots, but none of them point convincingly back to the central star. They have low-to-moderate ionisation with $T_ e $ sim 11000\,K e $ sim 1500 cm$^ $ and generally exhibit a higher extinction than the extended high-ionisation nebula. Within the MUSE field, a low-redshift emission-line galaxy was serendipitously found to be hiding behind NGC 4361. The spectrum of this dwarf galaxy was carefully extracted from the bright foreground nebular emission and the galaxy's line and continuum properties were then determined. NGC 4361 is not completely optically thin, as indicated by several extended regions and many compact features of lower ionisation emission. The low-ionisation 'freckles' identified here do not clearly appear to differ in (He, N, O, S) abundance with respect to the extended high-ionisation gas. The spatial distribution and radial velocities of these features suggest that they belong to a thick disk oriented perpendicular to the large-scale nebular gas, which may perhaps be remnants of an earlier structure. The low-luminosity disk galaxy at sim 87\,Mpc has bright H ii regions with metallicity 12+log(O/H) cong 8.4 and is suggested to be a Magellanic irregular or low-mass spiral.

Structural, electronic, and optical properties of Zn-doped V2O5 thin films

V2O5 shows a diverse range of applications due to its remarkable electronic and optical properties. This research is designed to tune the electronic and optical properties of V2O5 through modification in the energy band profile by varying Zn doping concentration. Density functional theory (DFT) calculations were used to investigate the Density of States (DOS) spectra for pure V2O5, exhibiting the prominent contribution of V-d and O-p orbitals, representing the p-d hybridized orbitals along with additional Zn-d orbital contribution in Zn-doped compositions. The effects of doping on the structural, morphological, elemental, and optical properties of the developed thin films were investigated employing x-ray diffraction (XRD), scanning electron microscope (SEM), x-ray dispersive spectroscopy (EDX), and spectroscopic ellipsometry (SE), respectively. x-ray diffraction analysis revealed the orthorhombic crystal structure in thin films. Surface morphology depicts the uniformly distributed compact rod-like features. The experimentally calculated band gap was found to decrease with Zn doping from 2.77 eV for pure V2O5 to 2.45 eV for maximum doping content. A significant variation is recorded in optical parameters like the increase in absorption coefficient and optical conductivity, which makes these more favorable for optoelectronic devices, particularly focusing on photovoltaics.

Design of an Improved Model for Lumpy Skin Disease Prediction Using Graph Neural Networks and Multiple Modal Data Fusion

An effective Lumpy Skin Disease prediction system mitigates economic losses and ensures animal health. Conventional approaches always fall short in terms of spatial-temporal analytics, very oft coming from a single source of data samples. To this regard, this research addresses the limitations by proposing a holistically integrated approach that puts together advanced machine learning techniques with enhanced disease prediction and management. In this paper, we provide a new framework that integrates graph neural networks for the fusion of graph-structured data with multivariables, attention mechanism, and self-supervised learning into compact feature representations. First, graph convolutional neural networks are utilized for spatial-temporal analysis. GCNs use node features, for instance, epidemiological data, and edge features in the form of connectivity data to generate node embeddings that capture complex dynamics of disease spread. It improved the prediction accuracy by up to 85%, greatly improving the spatial understanding in LSD dynamics. It basically fuses satellite image data, weather data, and livestock movement records through a multiple modal data fusion approach based on multiple head attention networks according to these embeddings. This will result in an enriched feature space in which the attention weights give insights into the relative importance of each data modality. This step improved the prediction accuracy by a further 5%, making the total prediction accuracy to be at 90%. Further fine-tuning of the model is done using self-supervised learning techniques through contrastive learning: SimCLR. This pre-paid feature representations using a large data set with unlabeled historic records on animal health data and environmental data samples. It improves model performance by up to 10% when fine-tuning the representations using LSD-specific data, thereby reducing reliance on labeled data and significantly improving the robustness and scalability of the model. This resulted in an overall accuracy of 95% for the fusion model, attained from the combination of these state-of-the-art techniques. In view of such comprehensive and diversified feature space, it showed very high resilience to adversarial attacks. Veterinary applications, in particular, support interpretability through attention mechanisms and node embeddings. Therefore, this work offers an accurate, robust, and interpretable solution to predict and manage Lumpy Skin Disease, contributing seriously to the domain of veterinary epidemiology.

Optically enhanced organic phototransistors for adaptive image processing under complex light conditions

Effectively handling massive information in complex lighting conditions is crucial for security-focused machine vision, as it relies on precise image feature detection and encryption/decryption to ensure integrity and confidentiality. Traditional hardware, primarily based on complementary metal-oxide-semiconductor (CMOS) technology, struggles with the complexity of processing tasks around the clock. To address this challenge, a ternary organic heterostructure visuomorphic phototransistor was tailored for adaptive critical image processing. The light coupling induced by the porous heterostructural layer enhances the light absorption efficiency by 2.5-fold. Our phototransistor exhibits bidirectional photoresponse across the visible and near-infrared (NIR) spectral regions, enabling adaptive edge detection in varying lighting conditions with precision above 85 %. The phototransistor array facilitates image encryption/decryption with a superior accuracy of 12 %/90 % compared to existing counterparts. This ternary organic hetero-integration combines multiple response modes, enabling compact and efficient critical feature processing.

Ultra-low-energy operation of electromagnetic bi-stable actuator under restricted energy supply

During the periodic drug administration for chronic diseases, unexpected battery depletion can be significantly problematic for the patients. While body heat harvesting utilizing the wearable thermoelectric generator (WTEG) shows the significant potential in continuous drug delivery systems (DDSs) by providing ceaseless power source, the power and the voltage output might not be sufficient to drive micropump actuators in such drug delivery systems. The continuous actuating current also imposes an additional challenge in developing efficient and compact voltage boosting circuits. In this paper, we propose an ultra-low-energy operating method of the WTEG-powered electromagnetic micropump system. In combination with our electromagnetic bi-stable actuator (EBA) where the energy demand is discretized, we present (1) the dedicated high-efficient charge pumps (CPs) (2) and the actuating pulse control method to prevent redundant energy waste. The proposed 10-stage charge pump is designed to reduce the conversion loss by utilizing multiple flying-capacitor switching structure, and the 2-stage series–parallel charge pump (SPCP) is fabricated with flexible thin-film super-capacitors (TFSCs) to realize the high boosting performance within compact and wearable features. Applying the proposed control method with the 10-stage CP in our WTEG-powered micropump system, we significantly reduce the charging time per actuation by 88.8% compared to the commercial boosting circuit. Also, the 2-stage SPCP achieves 84.0% reduction in the charging time while utilizing flexible and wearable features, showing a significant potential for self-powered wearable healthcare applications.

Generative Denoise Distillation: Simple stochastic noises induce efficient knowledge transfer for dense prediction

Knowledge distillation is the process of transferring knowledge from a more powerful large model (teacher) to a simpler counterpart (student). Numerous current approaches involve the student imitating the knowledge of the teacher directly, which tend to learn each spatial location’s features indiscriminately. Real-world datasets frequently exhibit noise, motivating models to acquire compact and representative features instead of memorizing every details in the training data. To derive a more compact representation (concept feature) from the teacher, we propose an innovative method, termed Generative Denoise Distillation (GDD), where stochastic noises are added to the concept feature of the student to embed them into the generated instance feature from a shallow network. Then, the generated instance feature is aligned with the knowledge of the instance from the teacher. This process diminishes model redundancy while concurrently augmenting its capacity for generalization. We extensively experiment with object detection, instance segmentation, and semantic segmentation to demonstrate the versatility and effectiveness of our method. Notably, GDD achieves new state-of-the-art performance in the tasks mentioned above. We have achieved substantial improvements in semantic segmentation by enhancing PspNet and DeepLabV3, both of which are based on ResNet-18, resulting in mIoU scores of 74.67 and 77.69, respectively, surpassing their previous scores of 69.85 and 73.20 on the Cityscapes dataset of 20 categories. The code is available at https://github.com/ZhgLiu/GDD.

Enhancing speech emotion detection with Windowed Long-Term Average Spectrum and Logistic-Rectified Linear Unit

Speech emotion detection can help monitor people's mental health by analyzing their speech patterns and recognizing emotional indicators. The raw data is initially augmented using various techniques such as shifting, pitching, noise injection, and stretching. Some feature extraction methods may only produce a limited set of features, potentially missing out on important information present in the data. This can result in reduced accuracy and the inability to capture complex patterns and variations in speech signals. The planned work consists of two phases: attribute extraction and classification. In the first phase, the WLTAS (Windowed Long-Term Average Spectrum) feature extraction extracts spectral information from speech signals, providing a compact and representative feature representation. It captures the frequency content and energy distribution of speech over time, enabling the analysis of important spectral characteristics. In the second phase, the Logistic-Rectified Linear Unit (LoRLU) introduces sparsity. It enhances the discriminative power of the classifier by allowing only positive values to pass through while discarding negative values. The Gibbs Restricted Boltzmann Machine (GRBM) is a generative probabilistic model that studies a joint probability delivery of the input data. It is particularly effective in capturing dependencies and patterns in high-dimensional data. The performance analysis of WLTAS-L-GRBM (Windowed Long-Term Average Spectrum - Logistic-Rectified Linear Unit - Gibbs Restricted Boltzmann Machine) has been conducted using evaluation metrics such as accuracy (99%), precision (0.97), recall (1), F1 score (0.98), MSE (0.01), and AUC-ROC (0.97) and also evaluated against various speech emotion datasets.

MS23D: A 3D object detection method using multi-scale semantic feature points to construct 3D feature layer

LiDAR point clouds can effectively depict the motion and posture of objects in three-dimensional space. Many studies accomplish the 3D object detection by voxelizing point clouds. However, in autonomous driving scenarios, the sparsity and hollowness of point clouds create some difficulties for voxel-based methods. The sparsity of point clouds makes it challenging to describe the geometric features of objects. The hollowness of point clouds poses difficulties for the aggregation of 3D features. We propose a two-stage 3D object detection framework, called MS23D. (1) We propose a method using voxel feature points from multi-branch to construct the 3D feature layer. Using voxel feature points from different branches, we construct a relatively compact 3D feature layer with rich semantic features. Additionally, we propose a distance-weighted sampling method, reducing the loss of foreground points caused by downsampling and allowing the 3D feature layer to retain more foreground points. (2) In response to the hollowness of point clouds, we predict the offsets between deep-level feature points and the object’s centroid, making them as close as possible to the object’s centroid. This enables the aggregation of these feature points with abundant semantic features. For feature points from shallow-level, we retain them on the object’s surface to describe the geometric features of the object. To validate our approach, we evaluated its effectiveness on both the KITTI and ONCE datasets.