Pairs Of Features Research Articles

Crossmodal correspondences between visual features and tastes in preschoolers: an exploratory study.

Adults possess a natural inclination to associate sensory cues derived from distinct modalities, such as the pairing of sweet with pink. However, studies exploring crossmodal correspondences in children, particularly in the sensory pairing of visual features and tastes, are scant, leaving unanswered questions regarding the developmental trajectory of crossmodal correspondences. The present study investigates whether Japanese preschool children demonstrate specific biases in shape-color, shape-taste, and color-taste associations. In a series of in-person experiments, 92 children between 3 to 6 years of age completed matching tasks utilizing paper stimuli. Children exhibit crossmodal correspondences in shape-color (circle-red and asymmetrical star-yellow), shape-taste (triangle-salty and circle-sweet), and color-taste (yellow-sour, black-bitter, and pink-sweet) associations. Moreover, children's choices are not influenced by their individual preferences. The crossmodal correspondences observed in this study have been observed in previous research on adults from the same (Japanese) culture, although adults showed more crossmodal correspondences than the children in this study (e.g., pink-circle, triangle-sour, and green-bitter). Thus, while some crossmodal correspondences emerge during childhood, others may require additional time to develop, thereby highlighting the importance of understanding the cognitive mechanisms underlying crossmodal correspondences from an ontogenic perspective.

Intelligent assembly modeling of complex product based on cognition of interaction structures

Interactions of a part with other parts in an assembly constitute the behavior of the part. However, currently parts interact with each other lifelessly in assembly environment due to their incapability of cognizing interaction structures between them, making the assembly modeling of complex products suffer from considerable manual interventions. An intelligent assembly modeling method is proposed for complex products in this paper, by endowing parts with the ability of cognizing interaction structures between them. In this method, the concept of interaction feature pair (IFP) and the corresponding automatic construction algorithms are developed to explicitly describe interaction intent (i.e., how a part will mate with other parts) in designer’s mind, providing carriers for part cognition behaviors and dynamical interaction intent in the assembly modeling process. On this basis, an object-oriented product information model considering IFP is established to facilitate the complete capture of interaction intent at part modeling stage and the consistent transmission of it to assembly modeling stage, endowing parts with the cognition ability of foreseeing how to mate with other parts before and after loading into assembly modeling environment. Then, IFP matching algorithms are designed to support the part cognition behaviors of actively cognizing interaction structures, perceiving potential mating parts and mating with other parts with high efficiency. Assembly modeling case studies of a reducer and a Mecanum wheel with few manual interventions indicate the feasibility of the proposed approach.

SEGANet: 3D object detection with shape-enhancement and geometry-aware network

3D object detection approaches from point clouds develop rapidly. However, the distribution of point clouds is unbalanced in the real scene, and thus the distant or occluded objects suffer from too few points to be perceived. This case damages the overall detection accuracy. Hence, we propose a novel two-stage 3D object detection framework, the Shape-Enhancement and Geometry-Aware Network (SEGANet), which aims to mitigate the negative impact of unbalanced point distribution for boosting detection performance. In stage 1, we first capture fine-grained structural knowledge with the assistance of point-wise features from voxels to generate proposals. And in stage 2, we construct a shape enhancement module to reconstruct complete surface points for objects within proposals, then establish an elaborate geometric relevance-aware Transformer module to aggregate high-correlated feature pairs of reconstructed-known parts and decode vital geometric relations of aggregated features. Thus, critical geometric clues are supplied for objects from the data and feature levels, achieving enhanced features for box refinement. Extensive experiments on KITTI and Waymo datasets show that SEGANet achieves low model complexity and excellent detection accuracy, especially leading the baseline method by 2.18% gain in overall detection accuracy and 1.8% gain in average accuracy of weakly sensing objects. This verifies that SEGANet effectively alleviates the impact of point imbalance to significantly boost detection performance.

Boosting Few-Shot Fine-Grained Recognition With Background Suppression and Foreground Alignment

Few-shot fine-grained recognition (FS-FGR) aims to recognize novel fine-grained categories with the help of limited available samples. Undoubtedly, this task inherits the main challenges from both few-shot learning and fine-grained recognition. First, the lack of labeled samples makes the learned model easy to overfit. Second, it also suffers from high intra-class variance and low inter-class differences in the datasets. To address this challenging task, we propose a two-stage background suppression and foreground alignment framework, which is composed of a background activation suppression (BAS) module, a foreground object alignment (FOA) module, and a local-to-local (L2L) similarity metric. Specifically, the BAS is introduced to generate a foreground mask for localization to weaken background disturbance and enhance dominative foreground objects. The FOA then reconstructs the feature map of each support sample according to its correction to the query ones, which addresses the problem of misalignment between support-query image pairs. To enable the proposed method to have the ability to capture subtle differences in confused samples, we present a novel L2L similarity metric to further measure the local similarity between a pair of aligned spatial features in the embedding space. What's more, considering that background interference brings poor robustness, we infer the pairwise similarity of feature maps using both the raw image and the refined image. Extensive experiments conducted on multiple popular fine-grained benchmarks demonstrate that our method outperforms the existing state of the art by a large margin. The source codes are available at: https://github.com/CSer-Tang-hao/BSFA-FSFG.

Role of Simple Spatial Gradient in Reinforcing the Accuracy of Temperature Determination of HED Plasma via Spectral Line-Area Ratios

We report on the simulation of temperature gradients in tamped NaFMgO target-foil plasma, heated and backlit by z-pinch dynamic hohlraum radiation. Our approach compares the spectroscopic output of a collisional-radiative model (prismspect) with soft X-ray absorption spectra collected on Sandia National Laboratories’ (SNL) Z Pulsed Power Facility. The pattern of minimum χ2 is seen to agree with an efficient, three-parameter model. Results show that a negligible gradient in electron temperature Te is consistent with experimental data, justifying the assumptions of previous work. The predicted sensitivity of line spectra to the gradient-aligned profile of Te is documented for each spectral feature, so that the line-area ratio between a pair of spectral features may be assessed as a proxy for the existence and quantification of such gradients.

The binary structure of event files generalizes to abstract features: A nonhierarchical explanation of task set boundaries for the congruency sequence effect.

Current views posit that forming and retrieving memories of ongoing events influences action control. However, the organizational structure of these memories, or event files, remains unclear. The hierarchical coding view posits a hierarchical structure, wherein task sets occupy a high level of the hierarchy. Here, the contents of an event file can be retrieved only if the task set repeats. In contrast, the binary coding view posits a nonhierarchical structure, which consists of a collection of independent, binary bindings between different feature pairs. In this view, repeating an abstract feature from a previous event (e.g., the previous trial's S-R mapping) triggers the retrieval of the associated feature from the same binding (e.g., the previous trial's congruency) even if the task set changes. To distinguish between these views, we investigated the nature of task set boundaries for the congruency sequence effect (CSE), an index of adaptive control that reflects event file formation and retrieval. Specifically, we investigated whether or not a CSE appears when the task set changes but the previous trial's S-R mapping repeats. Three experiments involving a cross-modal prime-probe task yielded a CSE under these conditions and ruled out alternative explanations. These findings show that the typical binary structure of event files generalizes from concrete features (e.g., colors and locations) to abstract features (e.g., S-R mappings and task sets). Therefore, contrary to the hierarchical coding view, they provide a nonhierarchical explanation of task set boundaries for the CSE. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Edge Distraction-aware Salient Object Detection

Integrating low-level edge features has been proven to be effective in preserving clear boundaries of salient objects. However, the locality of edge features makes it difficult to capture globally salient edges, leading to distraction in the final predictions. To address this problem, we propose to produce distraction-free edge features by incorporating cross-scale holistic interdependencies between high-level features. In particular, we first formulate our edge features extraction process as a boundary-filling problem. In this way, we enforce edge features to focus on closed boundaries instead of those disconnected background edges. Secondly, we propose to explore cross-scale holistic contextual connections between every position pair of high-level feature maps regardless of their distances across different scales. It selectively aggregates features at each position based on its connections to all the others, simulating the “contrast” stimulus of visual saliency. Finally, we present a complementary features integration module to fuse low- and high-level features according to their properties. Experimental results demonstrate our proposed method outperforms previous state-of-the-art methods on the benchmark datasets, with the fast inference speed of 30 <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">FPS</i> on a single GPU.

An Underground Parking Lot Mask for the Long-Term Visual Localization of Intelligent Vehicles

In the underground parking lot (UPL) scenario, appearance changes in an indoor environment cause visual feature matching failures between features extracted from the image and features in the map, and further, lead to vehicle localization loss. In this article, a prior UPL mask is proposed for long-term vehicle localization. First, the UPL polygonal pattern is summarized based on structured environment characteristics. Then, feature matching is performed on the unannotated video data that reflect the degree of environmental change, and the polygonal pattern parameters are determined by a rule-based method using the statistical results of feature matching. Thus, the main distribution area of long-term static features in the mask is obtained quantitatively. Finally, the extraction of features in the mask for mapping and localization allows more static feature pairs to be matched for motion estimation. A real vehicle platform with stereo vision was set up for experimental verification. The comparison results demonstrated that the proposed method improved long-term vehicle localization stability and adaptability to overcome environmental changes.

Bidirectional feature matching based on deep pairwise contrastive learning for multiparametric MRI image synthesis

Objective. Multi-parametric MR image synthesis is an effective approach for several clinical applications where specific modalities may be unavailable to reach a diagnosis. While technical and practical conditions limit the acquisition of new modalities for a patient, multimodal image synthesis combines multiple modalities to synthesize the desired modality. Approach. In this paper, we propose a new multi-parametric magnetic resonance imaging (MRI) synthesis model, which generates the target MRI modality from two other available modalities, in pathological MR images. We first adopt a contrastive learning approach that trains an encoder network to extract a suitable feature representation of the target space. Secondly, we build a synthesis network that generates the target image from a common feature space that approximately matches the contrastive learned space of the target modality. We incorporate a bidirectional feature learning strategy that learns a multimodal feature matching function, in two opposite directions, to transform the augmented multichannel input in the learned target space. Overall, our training synthesis loss is expressed as the combination of the reconstruction loss and a bidirectional triplet loss, using a pair of features. Main results. Compared to other state-of-the-art methods, the proposed model achieved an average improvement rate of 3.9% and 3.6% on the IXI and BraTS’18 datasets respectively. On the tumor BraTS’18 dataset, our model records the highest Dice score of 0.793(0.04) for preserving the synthesized tumor regions in the segmented images. Significance. Validation of the proposed model on two public datasets confirms the efficiency of the model to generate different MR contrasts, and preserve tumor areas in the synthesized images. In addition, the model is flexible to generate head and neck CT image from MR acquisitions. In future work, we plan to validate the model using interventional iMRI contrasts for MR-guided neurosurgery applications, and also for radiotherapy applications. Clinical measurements will be collected during surgery to evaluate the model’s performance.

A random forest‐based fatigue damage classification framework using global loading and displacement information

AbstractThe article presents a random forest‐based damage classification methodology built upon the force and displacement data obtained via fatigue testing of Al7075‐T6 specimens. Four features, namely, average displacement (D), material stiffness (ST), energy dissipation rate (EDR), and cumulative energy dissipation (CED), are defined from the collected data. These features are classified into healthy and cracked states based on the information obtained from a high‐resolution confocal microscope and a digital microscope. Random forest‐based classifiers are, thereafter, trained and tested using the individual as well as different combinations of the four features. The results reveal that, when used individually, CED shows the best accuracy of 80% in detecting a fatigue crack. When two or more pairs of features are used, the accuracy increases significantly to beyond 95%. CED is observed to be the most important feature for developing efficient classifiers for fatigue crack detection. The results demonstrate that the random forest classifier is a viable option to accurately detect fatigue damage based on the four features using only global information.

Galaxy and mass assembly (GAMA): comparing visually and spectroscopically identified galaxy merger samples

ABSTRACT We conduct a comparison of the merging galaxy populations detected by a sample of visual identification of tidal features around galaxies as well as spectroscopically detected close pairs of galaxies to determine whether our method of selecting merging galaxies biases our understanding of galaxy interactions. Our volume-limited parent sample consists of 852 galaxies from the Galaxy And Mass Assembly (GAMA) survey in the redshift range 0.04 ≤ z ≤ 0.20 and stellar mass range 9.50 ≤ log $_{10}(M_{\star }/\rm {M}_{\odot })\le ~11.0$. We conduct our comparison using images from the Ultradeep layer of the Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP) to visually classify galaxies with tidal features and compare these to the galaxies in the GAMA spectroscopic close-pair sample. We identify 198 galaxies possessing tidal features, resulting in a tidal feature fraction ftidal = 0.23 ± 0.02. We also identify 80 galaxies involved in close pairs, resulting in a close pair fraction fpair = 0.09 ± 0.01. Upon comparison of our tidal feature and close pair samples we identify 42 galaxies that are present in both samples, yielding a fraction fboth = 0.05 ± 0.01. We find evidence to suggest that the sample of close pairs of galaxies is more likely to detect early stage mergers, where two separate galaxies are still visible, and the tidal feature sample detects later-stage mergers, where only one galaxy nucleus remains visible. The overlap of the close pair and tidal feature samples likely detect intermediate-stage mergers. Our results are in good agreement with the predictions of cosmological hydrodynamical simulations regarding the populations of merging galaxies detected by close pair and tidal feature samples.

Tumor immune infiltrates and overall survival in non-small cell lung cancer are associated with an unsupervised radiomic signature.

e20525 Background: We investigated whether radiomic features derived from computed tomography (CT) scans are predictive of survival and are correlated with immune cell types in non-small cell lung cancer (NSCLC), employing a mathematical network-based approach. Methods: DICOM data of pre-treatment CT scans and associated clinical variables for 116 NSCLC patients were downloaded from The Cancer Imaging Atlas (TCIA). RNA-Seq gene expression profiles for the matched cases were downloaded from the GEO database (GSE103584). CIBERSORT scores that consist of 22 immune cell types were generated using the R library (CIBERSORT) on the RNA-Seq gene expression profiles. From contoured tumors on CT scans, 83 radiomic features including tumor volume were extracted, utilizing the CERR radiomics toolbox. Radiomic networks were constructed with a threshold of Spearman correlation coefficient of 0.85 between each pair of radiomic features. The mathematical network-based k-means clustering method was applied to the largest connected network component that consists of 40 radiomic features, using the Wasserstein distance from optimal mass transport (OMT), to minimize the summed distance between k partitioned centroids and their respective data points. For the sub-groups identified, Kaplan-Meier survival analysis was conducted with log-rank test and CIBERSORT scores and clinical variables were then compared among the sub-groups using Kruskal-Wallis test. Results: The median age was 69 years (range: 46–85 years). Eighty-seven patients (75.0%) were males. Most patients were former (N = 75, 64.7%) or current smokers (N = 24, 20.7%). The median follow-up time (last follow-up or death) was 44 months and 75 patients were alive at last contact. The Wasserstein distance-based k-means method resulted in 3 sub-groups (high: N = 32, medium: N = 45, low-risk: N = 39). Kaplan-Meier analysis on the 3 sub-groups showed a significant difference in overall survival with log-rank p = 0.02. There were no significant differences for the variables mentioned above among the sub-groups. Two immune cell types showed statistically significant differences among the sub-groups: mast cells resting with p = 0.004 and M2 macrophages with p = 0.013 using Kruskal-Wallis test. The values for both immune cell types increased from high &lt; medium &lt; low-risk group. There was a significant volume difference with p &lt; 0.001, but the difference did not fully account for the survival difference; the median volume was 30.8 cc, 2.7 cc, and 12.8 cc for high, medium, and low-risk groups, respectively. Conclusions: A radiomic network analysis, utilizing a mathematical network-based method, resulted in sub-groups with the survival difference in NSCLC patients. Furthermore, these sub-groups had significantly different immune cell types likely associated with the survival.

Investigation on Features Extraction of Defect Signals Using Support Vector Machine and Multi-feature Fusion

This article investigated the sensitivity of each acoustic feature to steel plate defects through the coin-tapping method in order to improve acoustic detection methods precision and intelligent detection efficiency. 87 acoustic characteristic parameters of detection signals were researched including time domain, frequency domain, and time-frequency domain. 10 kinds of percussion conditions were designed to meet the two classifications of acoustic signals. The degree of contribution of acoustic parameters to damage detection was arranged by sensibility analysis, which could be used as a basis for selecting acoustic parameters for similar structural damage detection. The characteristic vectors of various parameter combinations were determined as the inputs of the acoustic detection method by analysing the sensitivity and correlation of different parameters. The multi-parameter feature vector with low redundancy was obtained by correlation analysis, which could be used as a combination method of the input value of the damage recognition classifier. The results show that the zero-crossing rate, main frequency and time-frequency domain parameters have high damage sensitivity, and there are always low-correlation feature pairs between them. The research results provide solutions for the selection and combination of feature parameters for similar data classification problems.

Inter-Frame Video Compression based on Adaptive Fuzzy Inference System Compression of Multiple Frame Characteristics

Video compression is used for storage or bandwidth efficiency in clip video information. Video compression involves encoders and decoders. Video compression uses intra-frame, inter-frame, and block-based methods. Video compression compresses nearby frame pairs into one compressed frame using inter-frame compression. This study defines odd and even neighboring frame pairings. Motion estimation, compensation, and frame difference underpin video compression methods. In this study, adaptive FIS (Fuzzy Inference System) compresses and decompresses each odd-even frame pair. First, adaptive FIS trained on all feature pairings of each odd-even frame pair. Video compression-decompression uses the taught adaptive FIS as a codec. The features utilized are "mean", "std (standard deviation)", "mad (mean absolute deviation)", and "mean (std)". This study uses all video frames' average DCT (Discrete Cosine Transform) components as a quality parameter. The adaptive FIS training feature and amount of odd-even frame pairings affect compression ratio variation. The proposed approach achieves CR=25.39% and P=80.13%. "Mean" performs best overall (P=87.15%). "Mean (mad)" has the best compression ratio (CR=24.68%) for storage efficiency. The "std" feature compresses the video without decompression since it has the lowest quality change (Q_dct=10.39%).

Virtual Reality Collision Detection Based on Improved Ant Colony Algorithm

In order to improve the performance in terms of detecting objects colliding in virtual reality, the ant colony algorithm was used to detect collisions. In the preliminary detection stage, the OBB bounding box and the spherical bounding box were used to detect the collision of objects, and the objects that may collide were selected. In the accurate detection stage, the model was sampled, and the feature pairs were used as the set to be detected for detecting collisions, the collision detection problem of the three-dimensional model was transformed into a nonlinear optimization problem of the distance between the feature pairs in the two-dimensional discrete space. The ant colony algorithm was introduced to solve the problem, and the pheromone concentration and update rules of the ant colony algorithm were optimized to improve the efficiency of the algorithm. The simulation results showed that, compared with the commonly used collision detection algorithms, our algorithm had high accuracy in detecting collisions and was less time-consuming.

Multi-source domain adaptation with spatio-temporal feature extractor for EEG emotion recognition

Electroencephalogram (EEG) emotion recognition suffers from cross-domain identification difficulties mainly caused by cross-subject and cross-session differences. Inspired by domain adaptation, the study constructs a novel multi-source domain adaptation with a Spatio-temporal feature extractor (MSDA-SFE) for EEG emotion recognition, which could reduce the signal differences between subjects and between sessions to make the target domain features align with the source domain features easily. In common feature learning, a Spatiotemporal feature extraction module is constructed to acquire domain-invariant characteristics of the source and target domains in which one subject is selected randomly as the target domain, and the remaining subjects are source domains. Whereafter, the obtained target domain features are paired with other source domain features respectively to generate N-1 pairs of concatenated features. And then these features are translated into N-1 domain-specific EEG features through N-1 parallel branches, where the target domain features are better aligned in the latent space, and the classification boundary of target samples is obtained by minimizing the disparity between branches. Finally, domain-specific features are fed to respective classifiers to get N-1 predictions, and their mean is put as the final result. Extensive experiments are implemented on SEED, SEED-IV, and DEAP benchmark databases. And, results indicate that the MSDA-SFE network outperforms the other comparison methods. Besides, the additional experiment also illustrates that the method has a powerful generalization capacity.

Bidirectional Association Discovery Leads to Precise Identification of Lung Cancer Biomarkers and Genome Taxa Class.

Identifying proximity between pairs of expression vectors is one of the fundamental requirements in machine learning and data mining algorithms. We propose a new metric, Bidirectional Association Similarity (BiAS), to measure the degree of mutual association between a pair of features and present a generalized formulation to compute BiAS between two vectors. Using non-linear programming optimization, we establish soundness of BiAS against the Jaccard and cosine similarities and prove that mutually associative features must be similar. The reverse, however, is not true. Finally, we show that BiAS is a transitive relation and can suitably be incorporated with any clustering algorithm, just like other metrics, to identify groups of mutually associative features in an ensemble. Experiments on clustering and classification of genome sequences for taxa identification and finding biomarkers in large airway epithelial cells expressions from smokers diagnosed with lung cancer reveal that knowledge precision is further improved with BiAS compared to seven other well-established metrics including the Pearson correlation coefficient, cosine similarity and the Jaccard similarity. Remarkably, the 10 out of the top 11 lung-cancer biomarkers found in the study using BiAS has been corroborated through previously reported clinically-backed studies. Thus, bidirectional association mining turns out effective for bio-knowledge discovery.

HandDGCL: Two‐hand 3D reconstruction based disturbing graph contrastive learning

AbstractVirtual reality (VR) and augmented reality (AR) applications are becoming increasingly prevalent. However, constructing realistic 3D hands, especially when two hands are interacting, from a single RGB image remains a major challenge due to severe mutual occlusion and the enormous diversity of hand poses. In this article, we propose a disturbing graph contrastive learning strategy for two‐hand 3D reconstruction. This involves a graph disturbance network designed to generate graph feature pairs to enhance the consistency of the two‐hand pose features. A contrastive learning module leverages high‐quality generative features for a strong feature expression. We further propose a similarity distinguish method to divide positive and negative features for accelerating the model convergence. Additionally, a multi‐term loss is designed to balance the relation among the hand pose, the visual scale and the viewpoint position. Our model has achieved state‐of‐the‐art results in the InterHand2.6M benchmark. Ablation studies show the model's great ability to correct unreasonable hand movements. In subjective assessments, our graph disturbance learning method significantly improves the construction of realistic 3D hands, especially when two hands are interacting.

A least square generative network based on invariant contrastive feature pair learning for multimodal MR image synthesis.

During MR-guided neurosurgical procedures, several factors may limit the acquisition of additional MR sequences, which are needed by neurosurgeons to adjust surgical plans or ensure complete tumor resection. Automatically synthesized MR contrasts generated from other available heterogeneous MR sequences could alleviate timing constraints. We propose a new multimodal MR synthesis approach leveraging a combination of MR modalities presenting glioblastomas to generate an additional modality. The proposed learning approach relies on a least square GAN (LSGAN) using an unsupervised contrastive learning strategy. We incorporate a contrastive encoder, which extracts an invariant contrastive representation from augmented pairs of the generated and real target MR contrasts. This contrastive representation describes a pair of features for each input channel, allowing to regularize the generator to be invariant to the high-frequency orientations. Moreover, when training the generator, we impose on the LSGAN loss another term reformulated as the combination of a reconstruction and a novel perception loss based on a pair of features. When compared to other multimodal MR synthesis approaches evaluated on the BraTS'18 brain dataset, the model yields the highest Dice score with [Formula: see text] and achieves the lowest variability information of [Formula: see text], with a probability rand index score of [Formula: see text] and a global consistency error of [Formula: see text]. The proposed model allows to generate reliable MR contrasts with enhanced tumors on the synthesized image using a brain tumor dataset (BraTS'18). In future work, we will perform a clinical evaluation of residual tumor segmentations during MR-guided neurosurgeries, where limited MR contrasts will be acquired during the procedure.

Interaction Metabolomics to Discover Synergists in Natural Product Mixtures.

Mass spectrometry metabolomics has become increasingly popular as an integral aspect of studies to identify active compounds from natural product mixtures. Classical metabolomics data analysis approaches do not consider the possibility that interactions (such as synergy) could occur between mixture components. With this study, we developed "interaction metabolomics" to overcome this limitation. The innovation of interaction metabolomics is the inclusion of compound interaction terms (CITs), which are calculated as the product of the intensities of each pair of features (detected ions) in the data matrix. Herein, we tested the utility of interaction metabolomics by spiking known concentrations of an antimicrobial compound (berberine) and a synergist (piperine) into a set of inactive matrices. We measured the antimicrobial activity for each of the resulting mixtures against Staphylococcus aureus and analyzed the mixtures with liquid chromatography coupled to high-resolution mass spectrometry. When the data set was processed without CITs (classical metabolomics), statistical analysis yielded a pattern of false positives. However, interaction metabolomics correctly identified berberine and piperine as the compounds responsible for the synergistic activity. To further validate the interaction metabolomics approach, we prepared mixtures from extracts of goldenseal (Hydrastis canadensis) and habañero pepper (Capsicum chinense) and correctly correlated synergistic activity of these mixtures to the combined action of berberine and several capsaicinoids. Our results demonstrate the utility of a conceptually new approach for identifying synergists in mixtures that may be useful for applications in natural products research and other research areas that require comprehensive mixture analysis.