Clean Data Research Articles

Knockoffs-SPR: Clean Sample Selection in Learning With Noisy Labels.

A noisy training set usually leads to the degradation of the generalization and robustness of neural networks. In this article, we propose a novel theoretically guaranteed clean sample selection framework for learning with noisy labels. Specifically, we first present a Scalable Penalized Regression (SPR) method, to model the linear relation between network features and one-hot labels. In SPR, the clean data are identified by the zero mean-shift parameters solved in the regression model. We theoretically show that SPR can recover clean data under some conditions. Under general scenarios, the conditions may be no longer satisfied; and some noisy data are falsely selected as clean data. To solve this problem, we propose a data-adaptive method for Scalable Penalized Regression with Knockoff filters (Knockoffs-SPR), which is provable to control the False-Selection-Rate (FSR) in the selected clean data. To improve the efficiency, we further present a split algorithm that divides the whole training set into small pieces that can be solved in parallel to make the framework scalable to large datasets. While Knockoffs-SPR can be regarded as a sample selection module for a standard supervised training pipeline, we further combine it with a semi-supervised algorithm to exploit the support of noisy data as unlabeled data. Experimental results on several benchmark datasets and real-world noisy datasets show the effectiveness of our framework and validate the theoretical results of Knockoffs-SPR.

Sparcle: Boosting the Accuracy of Data Cleaning Systems through Spatial Awareness

Though data cleaning systems have earned great success and wide spread in both academia and industry, they fall short when trying to clean spatial data. The main reason is that state-of-the-art data cleaning systems mainly rely on functional dependency rules where there is sufficient co-occurrence of value pairs to learn that a certain value of an attribute leads to a corresponding value of another attribute. However, for spatial attributes that represent locations, there is very little chance that two records would have the same exact coordinates, and hence co-occurrence is unlikely to exist. This paper presents Sparcle (SPatially-AwaRe CLEaning); a novel framework that injects spatial awareness into the core engine of rule-based data cleaning systems through two main concepts: (1) Spatial Neighborhood , where co-occurrence is relaxed to be within a certain spatial proximity rather than same exact value, and (2) Distance Weighting , where records are given different weights of whether they satisfy a dependency rule, based on their relative distance. Experimental results using a real deployment of Sparcle inside a state-of-the-art data cleaning system, and real and synthetic datasets, show that Sparcle significantly boosts the accuracy of data cleaning systems when dealing with spatial data.

The Optimization of LSTM Model by Wavelet Transform and Simulated Annealing Algorithm

The main objective of this paper is to study the integration of data processing methods and intelligent algorithms to optimize the results of LSTM prediction models. In this paper, continuous wavelet transform is used to clean and preprocess time series data and improve data quality. Wavelet reconstruction is used to restore the results. At the same time, the simulated annealing algorithm is introduced as an intelligent algorithm to search globally for the best solution to achieve the optimal prediction result. The application of this comprehensive approach can also improve the quality and precision of data analysis in various fields, such as the parameter estimation of pulse signals in physics. The core challenge of the research is to optimize the data prediction results, and for this purpose, a multi-level method of continuous wavelet transform, deep learning model (LSTM) and simulated annealing algorithm is adopted.

Lipid and sugar metabolism play an essential role in pollen development and male sterility: a case analysis in Brassica napus.

The genic male sterility (GMS) system is an important strategy for generating heterosis in plants. To better understand the essential role of lipid and sugar metabolism and to identify additional candidates for pollen development and male sterility, transcriptome and metabolome analysis of a GMS line of 1205AB in B. napus was used as a case study. To characterize the GMS system, the transcriptome and metabolome profiles were generated for 24 samples and 48 samples of 1205AB in B. napus, respectively. Transcriptome analysis yielded a total of 156.52 Gb of clean data and revealed the expression levels of 109,541 genes and 8,501 novel genes. In addition, a total of 1,353 metabolites were detected in the metabolomic analysis, including 784 in positive ion mode and 569 in negative ion mode. A total of 15,635 differentially expressed genes (DEGs) and 83 differential metabolites (DMs) were identified from different comparison groups, most of which were involved in lipid and sugar metabolism. The combination of transcriptome and metabolome analysis revealed 49 orthologous GMS genes related to lipid metabolism and 46 orthologous GMS genes related to sugar metabolism, as well as 45 novel genes. The transcriptome and metabolome profiles and their analysis provide useful reference data for the future discovery of additional GMS genes and the development of more robust male sterility breeding systems for use in the production of plant hybrids.

A self‐supervised scheme for ground roll suppression

AbstractIn recent years, self‐supervised procedures have advanced the field of seismic noise attenuation, due to not requiring a massive amount of clean labelled data in the training stage, an unobtainable requirement for seismic data. However, current self‐supervised methods usually suppress simple noise types, such as random and trace‐wise noise, instead of the complicated, aliased ground roll. Here, we propose an adaptation of a self‐supervised procedure, namely, blind‐fan networks, to remove aliased ground roll within seismic shot gathers without any requirement for clean data. The self‐supervised denoising procedure is implemented by designing a noise mask with a predefined direction to avoid the coherency of the ground roll being learned by the network while predicting one pixel's value. Numerical experiments on synthetic and field seismic data demonstrate that our method can effectively attenuate aliased ground roll.

Breaking Through the Noisy Correspondence: A Robust Model for Image-Text Matching

Unleashing the power of image-text matching in real-world applications is hampered by noisy correspondence. Manually curating high-quality datasets is expensive and time-consuming, and datasets generated using diffusion models are not adequately well-aligned. The most promising way is to collect image-text pairs from the Internet, but it will inevitably introduce noisy correspondence. To reduce the negative impact of noisy correspondence, we propose a novel model that first transforms the noisy correspondence filtering problem into a similarity distribution modeling problem by exploiting the powerful capabilities of pre-trained models. Specifically, we use the Gaussian Mixture model to model the similarity obtained by CLIP as clean distribution and noisy distribution, to filter out most of the noisy correspondence in the dataset. Afterward, we used relatively clean data to fine-tune the model. To further reduce the negative impact of unfiltered noisy correspondence, i.e., a minimal part where two distributions intersect during the fine-tuning process, we propose a distribution-sensitive dynamic margin ranking loss, further increasing the distance between the two distributions. Through continuous iteration, the noisy correspondence gradually decreases and the model performance gradually improves. Our extensive experiments demonstrate the effectiveness and robustness of our model even under high noise rates.

Sub-Band Backdoor Attack in Remote Sensing Imagery

Remote sensing datasets usually have a wide range of spatial and spectral resolutions. They provide unique advantages in surveillance systems, and many government organizations use remote sensing multispectral imagery to monitor security-critical infrastructures or targets. Artificial Intelligence (AI) has advanced rapidly in recent years and has been widely applied to remote image analysis, achieving state-of-the-art (SOTA) performance. However, AI models are vulnerable and can be easily deceived or poisoned. A malicious user may poison an AI model by creating a stealthy backdoor. A backdoored AI model performs well on clean data but behaves abnormally when a planted trigger appears in the data. Backdoor attacks have been extensively studied in machine learning-based computer vision applications with natural images. However, much less research has been conducted on remote sensing imagery, which typically consists of many more bands in addition to the red, green, and blue bands found in natural images. In this paper, we first extensively studied a popular backdoor attack, BadNets, applied to a remote sensing dataset, where the trigger was planted in all of the bands in the data. Our results showed that SOTA defense mechanisms, including Neural Cleanse, TABOR, Activation Clustering, Fine-Pruning, GangSweep, Strip, DeepInspect, and Pixel Backdoor, had difficulties detecting and mitigating the backdoor attack. We then proposed an explainable AI-guided backdoor attack specifically for remote sensing imagery by placing triggers in the image sub-bands. Our proposed attack model even poses stronger challenges to these SOTA defense mechanisms, and no method was able to defend it. These results send an alarming message about the catastrophic effects the backdoor attacks may have on satellite imagery.

A non-targeted analysis workflow for the identification of organic contaminants in a sludge water based on fragmentation matching score and metadata

ABSTRACT Non-targeted analysis (NTA) methods have been increasingly used to discover organic contaminants in environmental samples. The identification efficiency of conventional NTA workflow is often very low due to many factors such as low fragmentation score between experimental MS2 data and reference MS2 data/in-silico MS2 data. Additionally, the prioritisation of possible candidates is also challenging due to the large number of candidates obtained from MS acquisitions. In this study, a multiplatform data processing workflow applicable to non-targeted LC-MS data is proposed to improve the identification accuracy of organic contaminants in water. This workflow consists of (1) scanning quadrupole data-independent acquisition (SONAR) to acquire relatively clean MS1 and MS2 data, (2) data deconvolution using Progenesis QI Software to filter/select the appropriate ions (MS1 data), (3) the selected MS1 data matching against U.S. EPA Chemistry Dashboard to get a list of the environmentally related candidates with metadata (e.g., data sources, toxicity, and exposure), (4) experimental MS2 data matching against in-silico fragmentation data of the obtained candidate by MetFrag to obtain a fragmentation score, (5) compound identification through ranking the candidates by MetFrag based on fragmentation score and metadata. The identification efficiency of the metadata-based NTA workflow was evaluated using a 33-component mixture of chemicals in a complex sludge water. Out of 33 spiked compounds, 26 compounds had been identified using the defined NTA workflow. Most of the 26 identified compounds were recognized as 1st hit of the candidates obtained from the library searching. Overall, the incorporation of fragmentation score and metadata to rank the candidates has significantly improved the accuracy of the organic contaminant identification in the sludge water, compared to the conventional NTA workflow based solely on the fragmentation score.

ORDerly: Data Sets and Benchmarks for Chemical Reaction Data.

Machine learning has the potential to provide tremendous value to life sciences by providing models that aid in the discovery of new molecules and reduce the time for new products to come to market. Chemical reactions play a significant role in these fields, but there is a lack of high-quality open-source chemical reaction data sets for training machine learning models. Herein, we present ORDerly, an open-source Python package for the customizable and reproducible preparation of reaction data stored in accordance with the increasingly popular Open Reaction Database (ORD) schema. We use ORDerly to clean United States patent data stored in ORD and generate data sets for forward prediction, retrosynthesis, as well as the first benchmark for reaction condition prediction. We train neural networks on data sets generated with ORDerly for condition prediction and show that data sets missing key cleaning steps can lead to silently overinflated performance metrics. Additionally, we train transformers for forward and retrosynthesis prediction and demonstrate how non-patent data can be used to evaluate model generalization. By providing a customizable open-source solution for cleaning and preparing large chemical reaction data, ORDerly is poised to push forward the boundaries of machine learning applications in chemistry.

Integrated analysis of the gonadal methylome and transcriptome provides new insights into the expression regulation of sex determination and differentiation genes in spotted scat (Scatophagus argus)

The spotted scat (Scatophagus argus) is an economically important mariculture fish. There have been no studies of gonadal DNA methylation at the genomic level in S. argus. Here, the gonadal methylome profiles of male and female S. argus were constructed and a link between DNA methylation and transcriptional regulation was identified. The gonads at stage IV of three male and female S. argus individuals were subjected to whole-genome bisulfite sequencing (WGBS), generating an average of 17.76 Gb of clean data for each sample. Analysis of methylation levels in different genomic regions revealed that both male and female S. argus gonads exhibited low methylation levels in the promoters and first exons, but high levels in the inner exons and introns. In general, DNA methylation levels were higher in the testes than in the ovaries. A total of 234,482 differentially methylated regions (DMRs) were identified between testes and ovaries, including 234,245, 26 and 211 DMRs in the CpG, CHG, and CHH backgrounds, respectively. Combined methylation and transcriptome analysis revealed that the relationship between methylation patterns and gene expression was generally negatively correlated. The male-specific Dmrt1Y, which is the candidate sex-determining gene of S. argus, was lowly modified by DNA methylation at its promoter region in the testis. While its allele on the X chromosome, Dmrt1ΔX, was highly modified by DNA methylation in both the testis and ovary. The expression levels of some other genes involved in sex differentiation and sex steroid synthesis were also regulated by DNA methylation. This study sheds light on the divergent methylation patterns in the gonadal genomes of male and female S. argus. These findings laid the foundation for subsequent analysis of differentially expressed sex determination and differentiation genes between male and female S. argus, providing a crucial theoretical basis for sex-control breeding efforts in S. argus.

Dietary pattern and diversity analysis using DietDiveR in R: a cross-sectional evaluation in the National Health and Nutrition Examination Survey

BackgroundThere are few resources available for researchers aiming to conduct 24-h dietary record and recall analysis using R. ObjectivesWe aimed to develop DietDiveR, which is a toolkit of functions written in R for the analysis of recall or record data collected with the Automated Self-Administered 24-h Dietary Assessment Tool or 2-d 24-h dietary recall data from the National Health and Nutrition Examination Survey (NHANES). The R functions are intended for food and nutrition researchers who are not computational experts. MethodsDietDiveR provides users with functions to 1) clean dietary data, 2) analyze 24-h dietary intakes in relation to other study-specific metadata variables, 3) visualize percentages of energy intake from macronutrients, 4) perform principal component analysis or k-means clustering to group participants by similar data-driven dietary patterns, 5) generate foodtrees based on the hierarchical food group information for food items consumed, 6) perform principal coordinate analysis taking food grouping information into account, and 7) calculate diversity metrics for overall diet and specific food groups. DietDiveR includes a self-paced tutorial on a website (https://computational-nutrition-lab.github.io/DietDiveR/). As a demonstration, we applied DietDiveR to a demonstration data set and data from NHANES 2015–2016 to derive a dietary diversity measure of nuts, seeds, and legumes consumption. ResultsAdult participants in the NHANES 2015–2016 cycle were grouped depending on the diversity in their mean consumption of nuts, seeds, and legumes. The group with the highest diversity in nuts, seeds, and legumes consumption had 3.8 cm lower waist circumference (95% confidence interval: 1.0, 6.5) than those who did not consume nuts, seeds, and legumes. ConclusionsDietDiveR enables users to visualize dietary data and conduct data-driven dietary pattern analyses using R to answer research questions regarding diet. As a demonstration of this toolkit, we explored the diversity of nuts, seeds, and legumes consumption to highlight some of the ways DietDiveR can be used for analyses of dietary diversity.

Analisis Penyediaan Air Bersih di Gedung Rektorat Universitas Tanjungpura

The provision of clean water in the Tanjungpura University Rectorate Building area until now has not been analyzed for its efficiency leve. In this research, 11 buildings including the Tanjungpura University Rectorate Building were reviewed. Therefore, it is very important to analyze whether the water available in these buildings is still sufficient, given that the need for clean water will continue to change from time to time. In this final research, clean water demand for the Tanjungpura University Rectorate Building area in 2023 is predicted, and clean water usage data is taken for 1 month of working days from March to April 2023. Tanjungpura University Rectorate Building Area itself has a total of 11 buildings with 2 types of areas, namely office areas and education areas. The results of observations of 5 meters in the Tanjungpura University Rectorate Building Area that the amount of clean water usage is 688.9 m3 / month. Then the analysis of clean water demand in the Tanjungpura University Rectorate Building Area is carried out both based on the building area and the total number of clean water users in the area. The results of this study found that the effectiveness of each area is known to be different, but in total, the amount of clean water usage is smaller than the predicted amount of clean water demand. So it can be concluded that the provision of clean water in the Tanjungpura University Rectorate Building area in 2023 is effective.

Cleaning and Harmonizing Medical Image Data for Reliable AI: Lessons Learned from Longitudinal Oral Cancer Natural History Study Data.

For deep learning-based machine learning, not only are large and sufficiently diverse data crucial but their good qualities are equally important. However, in real-world applications, it is very common that raw source data may contain incorrect, noisy, inconsistent, improperly formatted and sometimes missing elements, particularly, when the datasets are large and sourced from many sites. In this paper, we present our work towards preparing and making image data ready for the development of AI-driven approaches for studying various aspects of the natural history of oral cancer. Specifically, we focus on two aspects: 1) cleaning the image data; and 2) extracting the annotation information. Data cleaning includes removing duplicates, identifying missing data, correcting errors, standardizing data sets, and removing personal sensitive information, toward combining data sourced from different study sites. These steps are often collectively referred to as data harmonization. Annotation information extraction includes identifying crucial or valuable texts that are manually entered by clinical providers related to the image paths/names and standardizing of the texts of labels. Both are important for the successful deep learning algorithm development and data analyses. Specifically, we provide details on the data under consideration, describe the challenges and issues we observed that motivated our work, present specific approaches and methods that we used to clean and standardize the image data and extract labelling information. Further, we discuss the ways to increase efficiency of the process and the lessons learned. Research ideas on automating the process with ML-driven techniques are also presented and discussed. Our intent in reporting and discussing such work in detail is to help provide insights in automating or, minimally, increasing the efficiency of these critical yet often under-reported processes.

Credit Risk Assessment and Fraud Detection in Financial Transactions Using Machine Learning

Credit risk assessment and fraud detection are crucial tasks in the financial industry, vital to preserving financial organizations' legitimacy and sustainability. Traditional methods often fall short in accurately assessing risk and detecting fraudulent activities in a timely manner. In recent years, machine learning has emerged as a powerful tool for enhancing these processes, leveraging great dimensions of transactional statistics and superior algos for making more informed decisions. This research paper explores the usage of ML techniques in credit risk assessment and fraud detection within financial transactions.
 The paper begins with an overview of the importance of accurate risk assessment and fraud detection in financial transactions and introduces the role of machine learning in addressing these challenges. A comprehensive literature review is conducted to analyze existing methodologies, algorithms, and research trends in the field. Data acquisition and preprocessing techniques are discussed, emphasizing the importance of clean and relevant data for model training. Feature engineering strategies are explored to extract meaningful information from financial transaction data and enhance the predictive capabilities of machine learning models.
 Various machine learning algorithms suitable for credit risk assessment and fraud detection are examined, including LR, SVMs, RF, DTs and DNNs. The efficacy of these techniques is evaluated by discussing model metrics for assessment and ensemble approaches for boosting efficiency, with a focus on metrics such as accuracy, precision, recall, and ROC-AUC.
 The paper presents case studies and experimental results illustrating the application of machine learning models in real-world scenarios, highlighting their effectiveness in improving risk assessment and fraud detection processes. Additionally, difficulties such as imbalanced datasets, comprehensibility of the model and adherence to regulations are discussed, along with potential research directions and future trends in the field.
 In conclusion, this research emphasizes the transformative potential of machine learning in credit risk assessment and fraud detection within financial transactions. By leveraging advanced algorithms and data-driven approaches, financial institutions can enhance their decision-making processes, mitigate risks, and safeguard against fraudulent activities, ultimately contributing to a more secure and resilient financial ecosystem.

Improved Diagnostic Approach for BRB Detection and Classification in Inverter-Driven Induction Motors Employing Sparse Stacked Autoencoder (SSAE) and LightGBM

This study introduces an innovative approach to diagnostics, employing a unique combination of techniques including a stratified group K-fold cross-validation method and a sparse stacked autoencoder (SSAE) alongside LightGBM. By examining signatures derived from motor current, voltage, speed, and torque, the framework aims to effectively detect and classify broken rotor bars (BRBs) within inverter-fed induction machines. In this kind of cross-validation method, class labels and grouping factors are spread out across folds by distributing motor operational data attributes equally over target label stratification and extra grouping information. By integrating SSAE and LightGBM, a gradient-boosting framework, we elevate the precision and efficacy of defect diagnosis. The SSAE feature extraction algorithm proves to be particularly effective in identifying small BRB signatures within motor operational data. Our approach relies on comprehensive datasets collected from motor systems operating under diverse loading conditions, ranging from 0% to 100%. Using a sparse stacked autoencoder, the model lowers the dimensionality and noise of the motor fault data. It then sends the cleaned data to the LightGBM network for fault diagnosis. LightGBM leverages the attributes of the sparse stacked autoencoder to showcase the distinctive qualities associated with BRBs. This integration offers the potential to improve defect identification by furnishing input representations that are both more precise and more concise. The proposed model (SSAE with LightGBM) was trained using 80% of the data, while the remaining 20% was used for testing. To validate the proposed architecture, we evaluate the accuracy, precision, recall, and F1-scores of the results using motor global signals, with the help of confusion matrices with receiver operating characteristic (ROC) curves. Following the training of a new LightGBM model with refined hyperparameters through Bayesian optimization, we proceed to conduct the final classification utilizing the optimal feature subset. Evaluation of the test dataset indicates that the BRBs diagnostic framework facilitates the detection and classification of issues with induction motor BRBs, achieving accuracy rates of up to 99% across all loading conditions.

HP3D-V2V: High-Precision 3D Object Detection Vehicle-to-Vehicle Cooperative Perception Algorithm.

Cooperative perception in the field of connected autonomous vehicles (CAVs) aims to overcome the inherent limitations of single-vehicle perception systems, including long-range occlusion, low resolution, and susceptibility to weather interference. In this regard, we propose a high-precision 3D object detection V2V cooperative perception algorithm. The algorithm utilizes a voxel grid-based statistical filter to effectively denoise point cloud data to obtain clean and reliable data. In addition, we design a feature extraction network based on the fusion of voxels and PointPillars and encode it to generate BEV features, which solves the spatial feature interaction problem lacking in the PointPillars approach and enhances the semantic information of the extracted features. A maximum pooling technique is used to reduce the dimensionality and generate pseudoimages, thereby skipping complex 3D convolutional computation. To facilitate effective feature fusion, we design a feature level-based crossvehicle feature fusion module. Experimental validation is conducted using the OPV2V dataset to assess vehicle coperception performance and compare it with existing mainstream coperception algorithms. Ablation experiments are also carried out to confirm the contributions of this approach. Experimental results show that our architecture achieves lightweighting with a higher average precision (AP) than other existing models.

From Toxic to Trustworthy: Using Self-Distillation and Semi-supervised Methods to Refine Neural Networks

Despite the tremendous success of deep neural networks (DNNs) across various fields, their susceptibility to potential backdoor attacks seriously threatens their application security, particularly in safety-critical or security-sensitive ones. Given this growing threat, there is a pressing need for research into purging backdoors from DNNs. However, prior efforts on erasing backdoor triggers not only failed to withstand increasingly powerful attacks but also resulted in reduced model performance. In this paper, we propose From Toxic to Trustworthy (FTT), an innovative approach to eliminate backdoor triggers while simultaneously enhancing model accuracy. Following the stringent and practical assumption of limited availability of clean data, we introduce a self-attention distillation (SAD) method to remove the backdoor by aligning the shallow and deep parts of the network. Furthermore, we first devise a semi-supervised learning (SSL) method that leverages ubiquitous and available poisoned data to further purify backdoors and improve accuracy. Extensive experiments on various attacks and models have shown that our FTT can reduce the attack success rate from 97% to 1% and improve the accuracy of 4% on average, demonstrating its effectiveness in mitigating backdoor attacks and improving model performance. Compared to state-of-the-art (SOTA) methods, our FTT can reduce the attack success rate by 2 times and improve the accuracy by 5%, shedding light on backdoor cleansing.

IOFM: Using the Interpolation Technique on the Over-Fitted Models to Identify Clean-Annotated Samples

Most recent state-of-the-art algorithms for handling noisy label problems are based on the memorization effect, which is a phenomenon that deep neural networks (DNNs) memorize clean data before noisy ones. While the memorization effect can be a powerful tool, there are several cases where memorization effect does not occur. Examples are imbalanced class distributions and heavy contamination on labels. To address this limitation, we introduce a whole new approach called the interpolation with the over-fitted model (IOFM), which leverages over-fitted deep neural networks. The IOFM utilizes a new finding of over-fitted DNNs: for a given training sample, its neighborhoods chosen from the feature space are distributed differently on the original input space depending on the cleanness of the target sample. The IOFM has notable features in two aspects: 1) it yields superior results even when the training data are imbalanced or heavily noisy, 2) since we utilize over-fitted deep neural networks, a fine-tuning procedure to select the optimal training epoch, which is an essential yet sensitive factor for the success of the memorization effect, is not required, and thus, the IOFM can be used for non-experts. Through extensive experiments, we show that our method can serve as a promising alternative to existing solutions dealing with noisy labels, offering improved performance even in challenging situations.

Abstract 7636: Quantitative analysis of colorectal adenocarcinoma images obtained by single-shot 17-plex staining followed by imaging with the Orion™ spatial biology platform

Abstract Background To understand the tumor microenvironment (TME), high resolution imaging of multiple biomarkers with whole slide context can be used as a basis for downstream biomarker quantitation and predicting patient outcomes1. Here we investigate a sample of invasive colorectal adenocarcinoma using whole slide, single-step high-plex staining and imaging at single-cell resolution followed by quantitative analysis. Methods The slide was stained in one round with a 17-plex immune-oncology panel then imaged in one round with the Orion™ spatial biology platform. Quantitative analysis was performed on the resulting TIFF file using the following pipeline:- Manually annotate regions of interest (ROI) on the tissue- Segment whole slide images into cells with UnMICST and S3segmenter - Import mask and clean data in QuPath - Generate primary feature data table with QuPath - Classify cells into populations in QuPath - Generate spatial biomarkers - Compare cell populations and their spatial interactions in the different ROIs. Results Data revealed a distinction between normal colonic epithelium, well-differentiated adenocarcinoma with immune cell collection, and an infiltrating border of the carcinoma. Differences in immune cell content and spatial organization were measured, and the infiltrating border found to contrast with other tumor regions by showing a lower proliferative fraction (Ki-67, nuclear) and differences in E-cadherin and cytokeratin expression patterns. Conclusions These data highlight the importance of sufficient plex, resolution and whole slide context to derive reliable spatial biomarkers of potential prognostic value. Further, hours vs. days speed for whole-section multiplexed biomarker quantitation, along with same-section conventional chromogenic analysis, makes this approach suited to multi-patient clinical studies. 1 JR Lin et al., High-plex immunofluorescence imaging and traditional histology of the same tissue section for discovering image-based biomarkers, Nat Cancer, 4, 1036-1052 (2023). Citation Format: Edward Lo, Tad George, Selena Larkin. Quantitative analysis of colorectal adenocarcinoma images obtained by single-shot 17-plex staining followed by imaging with the Orion™ spatial biology platform [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7636.

Robust Deep Neural Network for Learning in Noisy Multi-Label Food Images.

Deep networks can facilitate the monitoring of a balanced diet to help prevent various health problems related to eating disorders. Large, diverse, and clean data are essential for learning these types of algorithms. Although data can be collected automatically, the data cleaning process is time-consuming. This study aims to provide the model with the ability to learn even when the data are not completely clean. For this purpose, we extend the Attentive Feature MixUp method to enable its learning on noisy multi-label food data. The extension was based on the hypothesis that during the MixUp phase, when a pair of images are mixed, the resulting soft labels should be different for each ingredient, being larger for ingredients that are mixed with the background because they are better distinguished than when they are mixed with other ingredients. Furthermore, to address data perturbation, the incorporation of the Laplace approximation as a post-hoc method was analyzed. The evaluation of the proposed method was performed on two food datasets, where a notable performance improvement was obtained in terms of Jaccard index and F1 score, which validated the hypothesis raised. With the proposed MixUp, our method reduces the memorization of noisy multi-labels, thereby improving its performance.