Morphological Feature Extraction Research Articles

Prediction of recurrence risk in endometrial cancer with multimodal deep learning.

Predicting distant recurrence of endometrial cancer (EC) is crucial for personalized adjuvant treatment. The current gold standard of combined pathological and molecular profiling is costly, hampering implementation. Here we developed HECTOR (histopathology-based endometrial cancer tailored outcome risk), a multimodal deep learning prognostic model using hematoxylin and eosin-stained, whole-slide images and tumor stage as input, on 2,072 patients from eight EC cohorts including the PORTEC-1/-2/-3 randomized trials. HECTOR demonstrated C-indices in internal (n = 353) and two external (n = 160 and n = 151) test sets of 0.789, 0.828 and 0.815, respectively, outperforming the current gold standard, and identified patients with markedly different outcomes (10-year distant recurrence-free probabilities of 97.0%, 77.7% and 58.1% for HECTOR low-, intermediate- and high-risk groups, respectively, by Kaplan-Meier analysis). HECTOR also predicted adjuvant chemotherapy benefit better than current methods. Morphological and genomic feature extraction identified correlates of HECTOR risk groups, some with therapeutic potential. HECTOR improves on the current gold standard and may help delivery of personalized treatment in EC.

Fine-grained recognition of bitter gourd maturity based on Improved YOLOv5-seg model

Bitter gourd, being perishable, requires timely harvesting. Delayed harvesting can result in a substantial reduction in fruit quality. while premature harvesting leads to underdeveloped fruit and decreased yields, the continuous flowering pattern in bitter gourd underscores the significance of accurately assessing fruit growth and ensuring timely harvesting for subsequent fruit setting and development. The current reliance on the experience of production personnel represents a substantial inefficiency. We present an improved real-time instance segmentation model based on YOLOv5-seg. The utilization of dynamic snake convolution enables the extraction of morphological features from the curved and elongated structure of bitter gourd. Diverse branch blocks enhance feature space diversity without inflating model size and inference time, contributing to improved recognition of expansion stages during bitter gourd growth. Additionally, the introduction of Focal-EIOU loss accurately locates the boundary box and mask, addressing sample imbalances in the L2 stage. Experimental results showcase remarkable accuracy rates of 99.3%, 93.8%, and 98.3% for L1, L2, and L3 stages using mAP@0.5. In comparison, our model outperforms other case segmentation models, excelling in both detection accuracy and inference speed. The improved YOLOv5-seg model demonstrates strong performance in fine-grained recognition of bitter gourd during the expansion stage. It efficiently segments bitter gourd in real-time under varying lighting and occlusion conditions, providing crucial maturity information. This model offers reliable insights for agricultural workers, facilitating precise harvesting decisions.

A novel hierarchical approach to insight to spectral characteristics in surface water of karst wetlands and estimate its non-optically active parameters using field hyperspectral data

Wetlands cover only around 6 % of the Earth's land surface, and are recognized as one of the three major ecosystems, alongside forests and oceans. The ecological structure and function of karst wetlands are unique due to the influence of geologic structure. At present, the unclear spectral morphology of surface water in karst wetlands poses a significant challenge in remote sensing estimation of non-optically active water quality parameters (NAWQPs). This study proposed a novel multi-scale spectral morphology feature extraction (MSFE) method to insight to spectral characteristics in surface water of karst wetlands, and further screen the sensitive features of NAWQPs. Then we constructed three remote sensing inversion strategies for NAWQPs (TN, TP, NH3_N, DO), including direct estimation, indirect estimation, and auxiliary estimation. Finally, we constructed a novel pH-based hierarchical analysis framework (pH_HA) to thoroughly explore the influence of alkalinity-biased characteristics of karst water on the spectral domain of NAWQPs and its estimation accuracy using in-situ hyperspectral data, respectively. We found that the spectral characteristics of karst waters at the first reflectance peak (580 nm) differed significantly from other water body types. The MSFE successfully captured the sensitive spectral domains for NAWQPs, and focused on between 500 and 600 nm and 900–960 nm. The sensitive features captured by MSFE improved estimation accuracy of NAWQPs (R2 >0.9). Direct estimation presented more stable performance compared to the auxiliary estimation (average RMSE of 0.366 mg/L), and the auxiliary estimation model further improved the retrieval accuracy of TN compared to direct estimation model (R2 increasing from 0.43 to 0.56). The novel hierarchical framework clearly revealed the notable changes in the sensitive spectral domains of NAWQPs under different pH values, and enabled more precise determination of spectral subdomains of NAWQPs, and identified the optimal spectral features. The pH_HA framework effectively improved the estimation accuracy of NAWQPs (R2 increased from 0.514 to over 0.9), and the estimation accuracies (R2) of four NAWQPs were all more than 0.9 when the pH value was over 8.5. Our works provide an effective approach for monitoring water quality in karst wetlands.

From video to vital signs: using personal device cameras to measure pulse rate and predict blood pressure using explainable AI

In this study, we address the need for low-cost, rapid, and reliable methods for blood pressure measurement to facilitate timely diagnosis and treatment of hypertension. We aim to enhance understanding and trust in the novel remote photoplethysmography technology, specifically as deployed by Lifelight®\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^\\circledR$$\\end{document}, for vital sign monitoring. Lifelight®\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^\\circledR$$\\end{document} utilizes remote photoplethysmography to extract physiologically meaningful features from video recordings on personal devices. This study describes the accurate calculation of pulse rates and the extraction of morphological and time series features from the remote photoplethysmography signal for blood pressure prediction. Unlike common approaches, distinct model types are employed for predicting systolic and diastolic blood pressure. Evaluation on an independent clinical laboratory’s dataset demonstrates the efficacy of the proposed method. The pulse rate accuracy, measured as root mean square error, is 1.9 beats per minute. The model for systolic blood pressure prediction shows a mean error of 1.2 mmHg and a standard deviation of 15.9 mmHg. The model for diastolic prediction shows a mean error of -\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$-$$\\end{document}3.2 mmHg and a standard deviation of 13.1 mmHg. This research establishes the validity of remote photoplethysmography technology by showcasing accurate pulse rate calculations and effective feature extraction for blood pressure prediction. The results indicate the potential for widespread adoption of remote photoplethysmography in measuring blood pressure and other vital signs.

Abstract 3726: Absence of synergistic effects for PARPi and immunotherapy combinations in ex vivo 3D micro tumor assay for ovarian cancer patients

Abstract Introduction: 50% of ovarian cancers are homologous recombination deficient (HRD), with 15% of the patients carrying a BRCA mutation. PARP inhibitors (PARPi) are designed to improve survival in HRD patients, but clinical responses to PARPi often do not match HRD status. In this study, a 3D ex vivo tumor testing platform was applied to quantify patient-specific drug responses to (combinations of) PARPi, chemo- and immunotherapies. Material and Methods: 81 patients with ovarian cancer eligible for chemotherapy were included in the trial between 2019 and 2022 in the Netherlands (IRB P18.032). Tumor clusters were isolated from ascites and solid tissue, seeded and exposed to three PARPi, standard-of-care chemotherapies and immunotherapies, both single agent and in combination. Ex vivo drug sensitivity was quantified using high-throughput 3D imaging and extraction of morphological features. Bliss scores were calculated for synergy analysis. Results: Results were generated in two weeks after sample collection. Ex vivo patient-specific response profiles were observed for all drug classes, with a small percentage of patients (7%) showing strong PARPi sensitivity reflected by strong tumor cluster shrinkage upon treatment. Significant correlation of drug sensitivity profiles was observed comparing the different PARPi between themselves, but also comparing PARPi and platinum responses. Interestingly, ex vivo sensitivity to olaparib and immunotherapy combinations demonstrated additive, but no synergistic effects (21 patients, 79 tests). In addition, olaparib responses were shown to be independent of immune responses (SEA). Conclusion: This study applied a predictive ex vivo 3D micro tumor testing platform to classify patient-specific sensitivity to (a combination of) PARPi, chemo- and immunotherapies in over 80 patients. The current dataset demonstrates the potential of ex vivo testing to identify patients that are responsive to PARPi. No synergistic effects were observed for these drugs in combination with immunotherapies. Discussion: Evaluation of PARPi responses in a functional 3D micro tumor assay could support identification of responsive patients prior to treatment. Distinct responses were observed for PARPi and immunotherapies. No synergistic but additive effects were measured, suggesting that ovarian cancer patients could benefit from PARPi and immunotherapies. Correlation of clinical and ex vivo drug responses to PARPi in combination with available HRD tests will further validate the platform for predicting patient clinical outcome. Citation Format: Lieke J. Ceton, Esmee Koedoot, Timothy J. Sijsenaar, Dieudonné J. van der Meer, Marta Montero Garcia, Jurrian van der Valk, Kristofer Maad, Cor de Kroon, Anne van Altena, Willemijn Vader, Judith Kroep, Nelleke Ottevanger. Absence of synergistic effects for PARPi and immunotherapy combinations in ex vivo 3D micro tumor assay for ovarian cancer patients [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 3726.

Simultaneous Pore Detection and Morphological Features Extraction in Laser Powder Bed Fusion with Image Processing.

Internal pore defects are inevitable during laser powder bed fusion (LPBF), which have a significant impact on the mechanical properties of the parts. Therefore, detecting pores and obtaining their morphology will contribute to the quality of LPBF parts. Currently, supervised models are used for defect image detection, which requires a large amount of LPBF sample data, image labeling, and computing power equipment during the training process, resulting in high detection costs. This study extensively collected LPBF sample data and proposed a method for pore defect classification by obtaining its morphological features while detecting pore defects in optical microscopy (OM) images under various conditions. Compared with other advanced models, the proposed method achieves better detection accuracy on pore defect datasets with limited data. In addition, quickly detecting pore defects in a large number of labeling ground truth images will also contribute to the development of deep learning. In terms of image segmentation, the average accuracy scores of this method in the test images exceed 85%. The research results indicate that the algorithm proposed in this paper is suitable for quickly and accurately identifying pore defects from optical microscopy images.

A fast hyperspectral change detection algorithm for agricultural crops based on low-rank matrix and morphological feature extraction

Crop change detection study is the foundation of agricultural sustainability. The inherent high spectral resolution of hyperspectral images, combined with multi-temporal datasets, facilitates the detection of subtle changes. To enhance the accuracy and applicability of hyperspectral change detection in agricultural scenes, this paper introduces a fast hyperspectral change detection approach for agricultural crops based on low-rank matrix and morphological feature extraction (FLRaMF). The goal is to improve detection precision and computational efficiency of the change detection process. The method initially employs rapid low-rank matrix extraction to separate changing and non-changing pixels in the spectral domain. Subsequently, spatial information is introduced using attribute profiles, restricting spectral anomalies through hyperspectral morphology, which ultimately improves the detection results. This study utilized four hyperspectral change detection datasets in agricultural crop scenarios, optimizing and analyzing parameters. Experimental results and analysis indicate that the FLRaMF method can achieve higher detection accuracy with reduced computation cost in unsupervised, default parameter scenarios when performing agricultural crop change detection tasks.

An Automated Prostate-cancer Prediction System (APPS) Based on Advanced DFO-ConGA2L Model using MRI Imaging Technique

The prostate cancer is a deadly form of cancer that assassinates a significant number of men due of its mediocre identification process. Images from people with cancer include important and intricate details that are difficult for conventional diagnostic methods to extract. This work establishes a novel Automated Prostate-cancer Prediction System (APPS) model for the goal of detecting and classifying prostate cancer utilizing MRI imaging sequences. The supplied medical image is normalized using a Coherence Diffusion Filtering (CDFilter) approach for improved quality and contrast. The appropriate properties are also extracted from the normalized image using the morphological and texture feature extraction approach, which helps to increase the classifier's accuracy. In order to train the classifier, the most important properties are also selected utilizing the cutting-edge Dragon Fly Optimized Feature Selection (DFO-FS) algorithm. Using this method greatly improves the classifier's overall disease diagnosis performance in less time and with faster processing. More specifically, the provided MRI input data are used to categorize the prostate cancer-affected and healthy tissues using the new Convoluted Gated Axial Attention Learning Model (ConGA2L) based on the selected features. This study compares and validates the performance of the APPS model by looking at several aspects using publicly available prostate cancer data.

Meet the authors: Hanchuan Peng, Peng Xie, and Feng Xiong

In a recent paper at Patterns, Hanchuan Peng, Peng Xie, and Feng Xiong from Southeast University describe a deep learning method to characterize complete single-neuron morphologies, which can discover neuron projection patterns of diverse cells and learn neuronal morphology representation. In this interview, the authors shared the story behind the paper and their research experience. This interview is a companion to these authors' recent paper, "DSM: Deep sequential model for complete neuronal morphology representation and feature extraction."1.

DSM: Deep sequential model for complete neuronal morphology representation and feature extraction

The full morphology of single neurons is indispensable for understanding cell types, the basic building blocks in brains. Projecting trajectories are critical to extracting biologically relevant information from neuron morphologies, as they provide valuable information for both connectivity and cell identity. We developed an artificial intelligence method, deep sequential model (DSM), to extract concise, cell-type-defining features from projections across brain regions. DSM achieves more than 90% accuracy in classifying 12 major neuron projection types without compromising performance when spatial noise is present. Such remarkable robustness enabled us to efficiently manage and analyze several major full-morphology data sources, showcasing how characteristic long projections can define cell identities. We also succeeded in applying our model to both discovering previously unknown neuron subtypes and analyzing exceptional co-expressed genes involved in neuron projection circuits.

Benggang segmentation via deep exchanging of digital orthophoto map and digital surface model features

Benggang is a typical fragmented erosional landscape in southern and southeastern China, posing significant risk to the local residents and economic development. Therefore, an efficient and accurate fine-grained segmentation method is crucial for monitoring the Benggang areas. In this paper, we propose a deep learning-based automatic segmentation method for Benggang by integrating high-resolution Digital Orthophoto Map (DOM) and Digital Surface Model (DSM) data. The DSM data is used to extract slope maps, aiming to capture primary morphological features. The proposed method consists of a dual-stream convolutional encoder-decoder network in which multiple cascaded convolutional layers and a skip connection scheme are used to extract morphological and visual features of the Benggang areas. The rich discriminative information in the DOM and slope data is fused by a channel exchanging mechanism that dynamically exchanges the most discriminative features from either the DOM or DSM stream according to their importance at the channel level. Evaluation experiments were conducted on a challenging dataset collected from Guangdong Province, China, and the results show that the proposed channel exchanging network based deep fusion method achieves 84.62% IoU in Benggang segmentation, outperforming several existing unimodal or multimodal baselines. The proposed multimodal segmentation method greatly improves the efficiency of large-scale discovery of Benggang, and thus is important for the management and restoration of Benggang in southern and southeastern China, as well as the monitoring of other similar erosional landscapes.

Deep learning based identification of aquatic and semi aquatic plants

AbstractBotanists use morphological features of leaves for aquatic and semi‐aquatic plants identification. In deep learning, the convolution process is efficacious to label images. Existing studies showed that, in deep learning‐based leaf identification, direct image pixel values are used. Leaf images have similar sizes and almost equal pixel values. So, the pixel‐based leaf naming procedure makes uncertainty in the feature results. The direct evaluation of image pixels for leaf image identification is not an effective method, because it is leading to a forgetting problem in continuous learning. Using the pre‐processed databases, researchers are achieving more than 99% accurate results using deep learning. However, the same model fails to reproduce the same result in databases. In this study, morphological features of aquatic and semi‐aquatic plant leaves are converted into digital descriptors to avoid uncertainty in the feature results. In this digital descriptor‐based deep learning method, morphological features of aquatic and semi‐aquatic plant leaves are used. This is equivalent to Botanists' morphological features‐based leaf identification technology. Morphological feature extraction and digital descriptor generation helped this model to achieve 95% accuracy. In this model, leaf morphological features are used for training, so this help to understand the leaf properties of other leaf databases.

From images to features: unbiased morphology classification via variational auto-encoders and domain adaptation

ABSTRACT We present a novel approach for the dimensionality reduction of galaxy images by leveraging a combination of variational auto-encoders (VAEs) and domain adaptation (DA). We demonstrate the effectiveness of this approach using a sample of low-redshift galaxies with detailed morphological type labels from the Galaxy Zoo Dark Energy Camera Legacy Survey (DECaLS) project. We show that 40-dimensional latent variables can effectively reproduce most morphological features in galaxy images. To further validate the effectiveness of our approach, we utilized a classical random forest classifier on the 40-dimensional latent variables to make detailed morphology feature classifications. This approach performs similar to a direct neural network application on galaxy images. We further enhance our model by tuning the VAE network via DA using galaxies in the overlapping footprint of DECaLS and Beijing-Arizona Sky Survey + Mayall z-band Legacy Survey, enabling the unbiased application of our model to galaxy images in both surveys. We observed that DA led to even better morphological feature extraction and classification performance. Overall, this combination of VAE and DA can be applied to achieve image dimensionality reduction, defect image identification, and morphology classification in large optical surveys.

PENet: A phenotype encoding network for automatic extraction and representation of morphological discriminative features

Abstract Digitalized natural history collections serve as vital ecological and evolutionary research resources. Specimen retrieval based on morphological features allows for the rapid acquisition of similar specimens from these collections, aiding in maximizing the utilization of their resources and catering to the requirements of related research. However, achieving this objective requires effective feature extraction and representation techniques. We developed a phenotype encoding network (PENet), a deep learning‐based model that combines hashing methods to automatically extract and encode discriminative features into hash codes. We evaluated the performance of PENet on six data sets, including a newly constructed beetle data set (6566 images), which covers over 60% of the genera within the six subfamilies of Scarabaeidae. Phenotype encoding network showed high performance in feature extraction and image retrieval, allowing users to input an image of a specimen and efficiently retrieve all specimens with similar morphology. Two visualization methods, t‐SNE and Grad‐CAM, were used to evaluate the representation ability of the hash codes. Additionally, by using the hash codes generated from PENet, a phenetic distance tree was constructed based on the beetle data set. The result indicated that the hash codes could reveal the phenetic distances and relationships among categories to a certain extent. PENet provides an automatic and efficient method to extract and represent morphological discriminative features. The generated hash code can be used as a low‐dimensional carrier of these features, enabling efficient specimen retrieval. Moreover, the distance information carried by these hash codes suggests their potential in systematics, deserving further exploration.

Implementing K-Nearest Neighbor to Classify Wild Plant Leaf as a Medicinal Plants

in leaf shape. Therefore, this study aimed to create a system to help increase public knowledge about wild plant leaves that also function as medicinal plants by the KNN method. Leaves of wild plants, namely Rumput Minjangan, Sambung Rambat, Rambusa, Brotowali, and Zehneria japonica, are also medicinal plants in comparison. Image processing techniques used were preprocessing, image segmentation, and morphological feature extraction. Preprocessing consists of scaling and splitting the RGB components and using an RGB component decomposition process to find the color component that best describes the leaf shape and generate the blue component image. The segmentation process used a thresholding technique with a gray threshold value (T) of less than 150, which best separates objects and backgrounds. Some morphological feature extraction used are area, perimeter, metric, eccentricity, and aspect ratio. Based on the results of this research, the KNN method with variations in K values, namely 13, 15, and 17, obtained a system accuracy of 94.44% with a total of 90% training data and 10% test data. This comparison also affected the increase in system accuracy.

Eyeblink Detection Algorithm Based on Joint Optimization of VME and Morphological Feature Extraction

Eyeblink Detection Algorithm Based on Joint Optimization of VME and Morphological Feature Extraction

Assessing the performance of fully supervised and weakly supervised learning in breast cancer histopathology

Fully supervised learning (FSL) and weakly supervised learning based on multiple instance learning (WSLMIL) have become two mainstream paradigms for performing computer-aided pathological diagnosis (CAPD). It is well known that the high-intensity annotation burden of FSL and the performance degradation due to poor training constraints of WSLMIL are stumbling blocks for clinical translation. Even more unfortunate is the lack of comprehensive experimental analysis to help researchers make content-specific trade-offs between FSL and WSLMIL. In this work, we systematically compare the performances of FSL and WSLMIL on lymph node metastasis in breast cancer using a publicly available dataset. By analyzing the results of 16 backbone networks in the FSL paradigm, we find that emerging networks based on transformer (PVTv2-B2) and multi-layer perceptron (CycleMLP-B3) are more advantageous for performing patch-level classification task than convolution-based structure (ResNet50); combining their output with morphological feature extraction can be better used to universally perform slide-level classification task. However, the slight improvement brought by the evolution of the backbone network may be overshadowed by the aggregation operation in 6 WSLMIL algorithms, whereas relying on the in-domain backbone network can achieve a stable and excellent prediction performance in both quantitative analysis and interpretability comparisons. All the experimental results ultimately illustrate that the combination of in-domain backbone network and emergent aggregation operation becomes an economical and efficient technical tool for CAPD, which can be regarded as a compromise between FSL and WSLMIL.

A novel CS-NET architecture based on the unification of CNN, SVM and super-resolution spectrogram to monitor and classify blood pressure using photoplethysmography

ContextContinuous blood pressure (BP) monitoring plays an important role while treating various cardiovascular diseases and hypertension. A high correlation between arterial blood pressure (ABP) and Photoplethysmogram (PPG) signal enables using a PPG signal to monitor and classify BP continuously. Control of BP in realtime is the basis for the prevention of hypertension. Proposed approachThis work proposes a CS-NET architecture by unifying CNN and SVM approaches to classify BP using PPG signals. The main objective of the CS-NET method is to establish an accurate and reliable algorithm for the ABP classification. MethodologyABP signals are labeled normal and abnormal using the hypertension criteria the American College of Cardiology (ACC)/American Heart Association (AHA) laid down. The proposed CS-NET model incorporates three critical steps in three successive stages. The first stage includes converting a preprocessed PPG signal into a time-frequency (TF) representation called a super-resolution spectrogram by superlet transform. The second stage uses a convolutional neural network (CNN) model with several hidden layers to extract morphological features from every PPG super-resolution spectrogram. The third stage uses a support vector machine (SVM) classifier to classify the PPG signal. ResultsPPG signals are used to train and test the proposed model. The performance of the proposed CS-NET method is tested using MIMIC-II, MIMIC-III, and PPG-BP-figshare database in terms of accuracy and F1 score. Moreover, the CS-NET method achieves better results with high accuracy when compared with other benchmark approaches that require an electrocardiogram signal for reference. ConclusionsThe proposed model achieved an aggregate classification accuracy of 98.21% across a five-fold cross-validation technique, making it a reliable approach for BP classification in clinical settings and realtime monitoring.

A deep learning approach for morphological feature extraction based on variational auto-encoder: an application to mandible shape

Shape measurements are crucial for evolutionary and developmental biology; however, they present difficulties in the objective and automatic quantification of arbitrary shapes. Conventional approaches are based on anatomically prominent landmarks, which require manual annotations by experts. Here, we develop a machine-learning approach by presenting morphological regulated variational AutoEncoder (Morpho-VAE), an image-based deep learning framework, to conduct landmark-free shape analysis. The proposed architecture combines the unsupervised and supervised learning models to reduce dimensionality by focusing on morphological features that distinguish data with different labels. We applied the method to primate mandible image data. The extracted morphological features reflected the characteristics of the families to which the organisms belonged, despite the absence of correlation between the extracted morphological features and phylogenetic distance. Furthermore, we demonstrated the reconstruction of missing segments from incomplete images. The proposed method provides a flexible and promising tool for analyzing a wide variety of image data of biological shapes even those with missing segments.

Classification of Character Types of Wayang Kulit Using Extreme Learning Machine Algorithm

Wayang Kulit, which is an original Indonesian culture, is conditioned by the meaning of life in every performance. However, Wayang Kulit is currently less popular among young people due to a lack of understanding of the art of Wayang Kulit performance. To be able to provide knowledge to the younger generation about Wayang Kulit, one of which is by introducing the characters that exist in Wayang Kulit performances. This study aims to build an image classification system for Wayang Kulit characters by applying the neural network method using Extreme Learning Machine (ELM) and morphological feature extraction. Morphological feature extraction provides information about the shape characteristics of objects present in the image which are then used for input in the classification process. The Extreme Learning Machine (ELM) method may arbitrarily establish the weight value between the input neurons and the hidden layer during the classification step, resulting in a quicker learning pattern. Based on the test results using the confusion matrix, the accuracy value is calculated to get a value of 81%.