Accuracy Of Convolutional Neural Network Research Articles

DMAF-NET: Deep Multi-Scale Attention Fusion Network for Hyperspectral Image Classification with Limited Samples

In recent years, deep learning methods have achieved remarkable success in hyperspectral image classification (HSIC), and the utilization of convolutional neural networks (CNNs) has proven to be highly effective. However, there are still several critical issues that need to be addressed in the HSIC task, such as the lack of labeled training samples, which constrains the classification accuracy and generalization ability of CNNs. To address this problem, a deep multi-scale attention fusion network (DMAF-NET) is proposed in this paper. This network is based on multi-scale features and fully exploits the deep features of samples from multiple levels and different perspectives with an aim to enhance HSIC results using limited samples. The innovation of this article is mainly reflected in three aspects: Firstly, a novel baseline network for multi-scale feature extraction is designed with a pyramid structure and densely connected 3D octave convolutional network enabling the extraction of deep-level information from features at different granularities. Secondly, a multi-scale spatial–spectral attention module and a pyramidal multi-scale channel attention module are designed, respectively. This allows modeling of the comprehensive dependencies of coordinates and directions, local and global, in four dimensions. Finally, a multi-attention fusion module is designed to effectively combine feature mappings extracted from multiple branches. Extensive experiments on four popular datasets demonstrate that the proposed method can achieve high classification accuracy even with fewer labeled samples.

Enhancing Brain Tumor Classification: A CNN-Based Approach with InceptionV3 and Xception

Brain tumors are among the most aggressive and deadly diseases, with a very short life expectancy at the highest grade. To combat this, early detection and treatment are crucial.In this approach, MRI images are used to analyze brain abnormalities. The manual investigation of brain tumor classification is a time-consuming task, and there might be possibilities of human errors. Hence, accurate analysis in a short span of time is essential. This approach presents the automatic brain tumor classification algorithm using a highly accurate Convolutional Neural Network (CNN), InceptionV3 and Xception algorithm for classification of Glioma, Meningioma and No tumor. The brain part is initially segmented by a thresholding approach followed by a morphological operation. The brain MRI is classified using CNN, Inceptionv3, and Xceptionv3 algorithms. The system's performance is evaluated using precision, recall, F1 score, and accuracy parameters.

Assessing Endoscopic Response in Locally Advanced Rectal Cancer Treated with Total Neoadjuvant Therapy: Development and Validation of a Highly Accurate Convolutional Neural Network.

Rectal tumors display varying degrees of response to total neoadjuvant therapy (TNT). We evaluated the performance of a convolutional neural network (CNN) in interpreting endoscopic images of either a non-complete response to TNT or local regrowth during watch-and-wait surveillance. Endoscopic images from stage II/III rectal cancers treated with TNT from 2012 to 2020 at a single institution were retrospectively reviewed. Images were labelled as Tumor or No Tumor based on endoscopy timing (before, during, or after treatment) and the tumor's endoluminal response. A CNN was trained using ResNet-50 architecture. The area under the curve (AUC) was analyzed during training and for two test sets. The main test set included images of tumors treated with TNT. The other contained images of local regrowth. The model's performance was compared to sixteen surgeons and surgical trainees who evaluated 119 images for evidence of tumor. Fleiss' kappa was calculated by respondent experience level. A total of 2717 images from 288 patients were included; 1407 (51.8%) contained tumor. The AUC was 0.99, 0.98, and 0.92 for training, main test, and local regrowth test sets. The model performed on par with surgeons of all experience levels for the main test set. Interobserver agreement was good ( = 0.71-0.81). All groups outperformed the model in identifying tumor from images of local regrowth. Interobserver agreement was fair to moderate ( = 0.24-0.52). A highly accurate CNN matched the performance of colorectal surgeons in identifying a noncomplete response to TNT. However, the model demonstrated suboptimal accuracy when analyzing images of local regrowth.

Creating realistic anterior segment optical coherence tomography images using generative adversarial networks

AimsTo develop a generative adversarial network (GAN) capable of generating realistic high-resolution anterior segment optical coherence tomography (AS-OCT) images.MethodsThis study included 142 628 AS-OCT B-scans from the American University of...

Deep learned features selection algorithm: Removal operation of anomaly feature maps (RO-AFM)

Class/Regression Activation Maps (CAMs/RAMs; AMs) are often embedded into Convolutional Neural Networks (CNNs) for checking activated regions on input images at estimation. CNNs sometime generate unreliable AMs, such as activated regions, are inappropriate. Because AM is calculated by stacking many feature maps generated by the final convolutional layer, when there are Anomaly Feature Maps (AFMs), unreliable AMs can be generated. For example, suppose we have a CNN that evaluates the heart. In this case, the feature maps that focus on regions unrelated to the heart (e.g., shoulders and esophagus) are AFMs. Additionally, we have a hypothesis that the estimation accuracy of CNNs is increased by removing AFMs. However, methods for automatically detecting and removing AFMs have not been sufficiently studied in previous research to improve the performance of CNNs. Therefore, we propose a method named “Removal Operation of Anomaly Feature Maps (RO-AFMs)” to automatically detect and remove AFMs. When applying an RO-AFM to the Global Average Pooling (GAP) feature vectors of a CNN, dimensions of the GAP vector are reduced. Therefore, an RO-AFM is regarded as a deep-feature selection algorithm. From the results of adopting an RO-AFM to a Regression CNN (R-CNN) for estimating pulmonary artery wedge pressure, which is one of the measurement score for representing cardiac anomaly state, improved reliability of AM and estimation accuracy were verified. A comparison of RO-AFM and the existing methods, i.e., Lasso and the Feature Selection Layer (FSL), indicated that RO-AFM performed slightly better on the estimation accuracy. The computation time required for RO-AFM to evaluate all features was 1.833 s on average, confirming that RO-AFM is a lightweight process. Therefore, RO-AFM is useful for constructing a medical CNN that emphasizes explainability (e.g., CNNs for estimating the risk of a disease or a test value from chest X-ray or computed tomography images).

Brain Tumor Detection Using Deep Learning

The brain tumor is one of the most dangerous, common and aggressive diseases which leads to a very short life expectancy at the highest grade. Thus, to prevent life from such disease, early recognition, and fast treatment is an essential step. In this approach, MRI images are used to analyze brain abnormalities. The manual investigation of brain tumor classification is a time-consuming task and there might have possibilities of human errors. Hence accurate analysis in a tiny span of time is an essential requirement. In this approach, the automatic brain tumor classification algorithm using a highly accurate Convolutional Neural Network (CNN) algorithm is presented. Initially, the brain part is segmented by thresholding approach followed by a morphological operation. The brain MRI is classified using CNN, Inceptionv3, and Xceptionv3 algorithms. The performance of the system is evaluated using precision, recall, F1 score and accuracy parameter

Evaluation of the effects of K-Means Clustered-Based Weight Quantization on a Keras Library Based Convolutional Neural Network for Hand Written Digit Image Recognition

A lot of Convolutional Neural Networks (CNNs) have been implemented using FPGAs for the past years. Subsequently, memory saving features were added to the CNN through weight quantization using K-means clustering. A future goal on an ASIC design, involving CNN and weight quantization working together in one chip, can give way to an automated procedure of memory-saving CNN design. In this paper an evaluation was done on the effect of quantizing the weights of a Keras library-based CNN using K means clustering. Various values of K in K-means clustering were tested to see its effects on the CNN accuracy performance. This paper presents first the design approach of a Keras library based Convolutional Neural Network (CNN) for hand-written digit images. It then presents a hardware model design of K-Means clustering algorithm using VHDL. The performance of CNN for image recognition was then tested for various levels of weight quantization using K-means clustering algorithm. Simulation results showed a compression of weights as high as 60% resulted to less than 1% reduction in CNN’s accuracy. The findings in this paper will serve as guide in determining the relevant values of K i.e. the compression ratio, for future ASIC design on this topic.

Comparison of Activation Functions in Convolutional Neural Network for Poisson Noisy Image Classification

Deep learning, specifically the Convolutional Neural Network (CNN), has been a significant technology tool for image processing and human health. CNNs, which mimic the working principles of the human brain, can learn robust representations of images. However, CNNs are susceptible to noise interference, which can impact classification performance. Choosing the right activation function can improve CNNs performance and accuracy. This research aims to test the accuracy of CNN with ResNet50, VGG16, and GoogleNet architectures combined with several activation functions such as ReLU, Leaky ReLU, Sigmoid, and Tanh in the classification of images that experience Poisson noise. Poisson noise is applied to each test data to evaluate CNN accuracy. The data used in this study consists of three scenarios of different numbers of classes, namely 3 classes, 5 classes, and 10 classes. The results showed that combining ResNet50 with the ReLU activation function produced the best performance in class recognition in each scenario of the number of classes experiencing Poisson noise interference. The model achieved 97% accuracy for 3-class data, 95% for 5-class data, and 90% for 10-class data. These results show that using ResNet50 with the ReLU activation function can provide excellent resistance to Poisson noise in image processing. It was found that as the number of classes increases, the accuracy of image recognition tends to decrease. This shows that the more complex the image classification task is with a larger number of classes, the more difficult it is for CNNs to distinguish between different classes. Doi: 10.28991/ESJ-2024-08-02-014 Full Text: PDF

Comparison of CNN and SVM Methods on Web-based Skin Disease Classification Process

Skin, as the outermost layer of the body, is often in contact with bacteria, germs and viruses because of its most external position. According to statistics from the 2009 Indonesian Health Profile, skin illness is the third most common ailment seen in outpatient settings across the country's hospitals. Therefore, maintaining healthy skin is important because it protects the body's internal organs from injury and attack by pathogens. The development of image classification, such as the classification of skin diseases, has become a focus in the health sector. This research analyses the performance of Convolutional Neural Network (CNN) and Support Vector Machine (SVM) in web-based skin disease classification and overcomes the problem of imbalanced training data. With data augmentation and preprocess, this research improves data generalization and compares performance metrics such as Recall, Accuracy, and F1 Score. The results show that the average accuracy of CNN is 83.8%, while SVM reaches 81%. Although both models have high metrics for the normal class, other more complicated classes can only be handled by CNN with a value of more than 0.9. Apart from that, the CNN method also provides a higher Confidence Score than SVM, as well as a faster execution time. In conclusion, the CNN method is superior and recommended for skin disease classification based on web applications based on various performance test results.

A SWOT ANALYSIS OF BREAST CANCER DIAGNOSIS IN DIGITAL MAMMOGRAPHY USING DEEP CONVOLUTIONAL NEURAL NETWORK

Breast cancer is a significant global health challenge that affects both men and women, leading to cause-specific deaths. Current early screening interventions, such as digital mammography (DM), are susceptible to high false-positives and false-negatives. This paper explores the potential of convolutional neural network (CNN), a form of artificial intelligence (AI), to support screening mammography with the aim to enhance accuracy in lesion detection, image classification and diagnostic prediction. Because the adoption of AI in cancer diagnosis is still in its infancy, the objective of this paper is to provide insight into the benefits and limitations of deep learning-based approaches to detect and diagnose cancer. An analysis of the implementation of CNN in AI-screening mammography models was conducted, using the SWOT strategic analysis tool. Internal strengths that improve the predictive accuracy of CNN include transfer learning and data augmentation, whereas the internal weaknesses include a lack of data standardisation and reproducibility. External opportunities consist of increased sensitivity in differentiating between microcalcifications and non-tumorous structures, improved predictive diagnosis and reduced workload. Nevertheless, integration within clinical settings must also consider the external threats of breaching patient privacy, automation biases and the role of clinical judgement.

Features in Backgrounds of Microscopy Images Introduce Biases in Machine Learning Analyses

Subvisible particles may be encountered throughout the processing of therapeutic protein formulations. Flow imaging microscopy (FIM) and backgrounded membrane imaging (BMI) are techniques commonly used to record digital images of these particles, which may be analyzed to provide particle size distributions, concentrations, and identities. Although both techniques record digital images of particles within a sample, FIM analyzes particles suspended in flowing liquids, whereas BMI records images of dry particles after collection by filtration onto a membrane. This study compared the performance of convolutional neural networks (CNNs) in classifying images of subvisible particles recorded by both imaging techniques. Initially, CNNs trained on BMI images appeared to provide higher classification accuracies than those trained on FIM images. However, attribution analyses showed that classification predictions from CNNs trained on BMI images relied on features contributed by the membrane background, whereas predictions from CNNs trained on FIM features were based largely on features of the particles. Segmenting images to minimize the contributions from image backgrounds reduced the apparent accuracy of CNNs trained on BMI images but caused minimal reduction in the accuracy of CNNs trained on FIM images. Thus, the seemingly superior classification accuracy of CNNs trained on BMI images compared to FIM images was an artifact caused by subtle features in the backgrounds of BMI images. Our findings emphasize the importance of examining machine learning algorithms for image analysis with attribution methods to ensure the robustness of trained models and to mitigate potential influence of artifacts within training data sets.

Deep learning performance compared to healthcare experts in detecting wrist fractures from radiographs: A systematic review and meta-analysis

ObjectiveTo perform a systematic review and meta-analysis of the diagnostic accuracy of deep learning (DL) algorithms in the diagnosis of wrist fractures (WF) on plain wrist radiographs, taking healthcare experts consensus as reference standard. MethodsEmbase, Medline, PubMed, Scopus and Web of Science were searched in the period from 1 Jan 2012 to 9 March 2023. Eligible studies were patients with wrist radiographs for radial and ulnar fractures as the target condition, studies using DL algorithms based on convolutional neural networks (CNN), and healthcare experts consensus as the minimum reference standard. Studies were assessed with a modified QUADAS-2 tool, and we applied a bivariate random-effects model for meta-analysis of diagnostic test accuracy data. ResultsOur study was registered at PROSPERO with ID: CRD42023431398. We included 6 unique studies for meta-analysis, with a total of 33,026 radiographs. CNN performance compared to reference standards for the included articles found a summary sensitivity of 92% (95% CI: 80%–97%) and a summary specificity of 93% (95% CI: 76%–98%). The generalized bivariate I-squared statistic indicated considerable heterogeneity between the studies (81.90%). Four studies had one or more domains at high risk of bias and two studies had concerns regarding applicability. ConclusionThe diagnostic accuracy of CNNs was comparable to that of healthcare experts in wrist radiographs for investigation of WF. There is a need for studies with a robust reference standard, external data-set validation and investigation of diagnostic performance of healthcare experts aided with CNNs. Clinical relevance statementDL matches healthcare experts in diagnosing WFs, which potentially benefits patient diagnosis.

Artificial intelligence applied to magnetic resonance imaging reliably detects the presence, but not the location, of meniscus tears: a systematic review and meta-analysis.

To review and compare the accuracy of convolutional neural networks (CNN) for the diagnosis of meniscal tears in the current literature and analyze the decision-making processes utilized by these CNN algorithms. PubMed, MEDLINE, EMBASE, and Cochrane databases up to December 2022 were searched in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) statement. Risk of analysis was used for all identified articles. Predictive performance values, including sensitivity and specificity, were extracted for quantitative analysis. The meta-analysis was divided between AI prediction models identifying the presence of meniscus tears and the location of meniscus tears. Eleven articles were included in the final review, with a total of 13,467 patients and 57,551 images. Heterogeneity was statistically significantly large for the sensitivity of the tear identification analysis (I2 = 79%). A higher level of accuracy was observed in identifying the presence of a meniscal tear over locating tears in specific regions of the meniscus (AUC, 0.939 vs 0.905). Pooled sensitivity and specificity were 0.87 (95% confidence interval (CI) 0.80-0.91) and 0.89 (95% CI 0.83-0.93) for meniscus tear identification and 0.88 (95% CI 0.82-0.91) and 0.84 (95% CI 0.81-0.85) for locating the tears. AI prediction models achieved favorable performance in the diagnosis, but not location, of meniscus tears. Further studies on the clinical utilities of deep learning should include standardized reporting, external validation, and full reports of the predictive performances of these models, with a view to localizing tears more accurately. Meniscus tears are hard to diagnose in the knee magnetic resonance images. AI prediction models may play an important role in improving the diagnostic accuracy of clinicians and radiologists. • Artificial intelligence (AI) provides great potential in improving the diagnosis of meniscus tears. • The pooled diagnostic performance for artificial intelligence (AI) in identifying meniscus tears was better (sensitivity 87%, specificity 89%) than locating the tears (sensitivity 88%, specificity 84%). • AI is good at confirming the diagnosis of meniscus tears, but future work is required to guide the management of the disease.

Dislocation descriptors of low and high angle grain boundaries with convolutional neural networks

Multiscale modeling of plasticity in polycrystalline metals is a long-standing challenge in part because of the lack of an accepted grain boundary descriptor, which has hindered the bridging of scales between atomistic simulations and meso-scale discrete dislocation dynamics (DDD) models. While grain boundary dislocations (GBDs) for low angle grain boundaries can be ascertained by Burgers circuit analyses, the dislocation structures of high angle grain boundaries have remained elusive because of overlapping dislocation core fields. Here, we use convolutional neural networks (CNNs) to establish the locations of GBDs responsible for the misorientations of <001> symmetrical-tilt Cu grain boundaries, from the local atomistic stress fields modeled with molecular dynamics (MD) simulations. We achieve accurate CNN predictions of GBDs with sub-angstrom resolution across an extensive set of low- to high-tilt-angle grain boundaries through a training dataset, generated from superposing continuum-representations of the MD stress fields of single dislocations that consider the contributions of both the Volterra and dislocation core fields. The approach paves the way for dislocation representations of the atomistic grain boundary structures modeled by MD simulations or density functional theory (DFT) calculations in mesoscale DDD models to elucidate multiscale plasticity effects.

A novel identification method of microseismic events based on empirical mode decomposition and artificial neural network features

Microseismic technology has been widely used in geological hazard monitoring. However, the effective identification and classification of microseismic events in mines have always been a challenge in hazard monitoring. In this paper, a new method combining the empirical mode decomposition (EMD) algorithm and artificial neural network is proposed for the classification and identification of microseismic waveforms. This is the first study to explore such a model in the microseismic monitoring of coal mines. The data for this study were collected from a coal mine in northern China. A total of 2768 microseismic events and 2435 non-microseismic events were manually selected from a large amount of data. Firstly, the microseismic waveform reconstruction was performed using the 4th to 8th Intrinsic Mode Function (IMF) obtained by the EMD algorithm. Secondly, the peak factor, clearance factor, impulse factor, kurtosis, skewness, and waveform factor of each reconstructed microseismic waveform were used as features. Subsequently, the convolutional neural network (CNN) was used to classify microseismic events. Among them, data from 3600 microseismic events were used as the training set, and data from 1603 microseismic events were used as the test set. Finally, the original waveform data were used as input and compared with the classification results processed with the EMD algorithm. The classification methods of the BP neural network and Radial Basis Function (RBF) neural network were used for verification. The results of the processed features showed a significant increase in accuracy. The classification accuracy of CNN, BP neural network, and RBF neural network are 97.37%, 91.99%, and 94.23%, respectively. The results show that the utilization of the feature processing technique and CNN algorithm in this study demonstrates superior efficacy in microseismic event classification, which can be used in practice.

A real-time and accurate convolutional neural network for fabric defect detection

As a practical and challenging task, deep learning-based methods have achieved effective results for fabric defect detection, however, most of them mainly target detection accuracy at the expense of detection speed. Therefore, we propose a fabric defect detection method called PEI-YOLOv5. First, Particle Depthwise Convolution (PDConv) is proposed to extract spatial features more efficiently while reducing redundant computations and memory access, reducing model computation and improving detection speed. Second, Enhance-BiFPN(EB) is proposed based on the structure of BiFPN to enhance the attention of spatial and channel feature maps and the fusion of information at different scales. Third, we improve the loss function and propose IN loss, which improves the problem that the original IOU loss is weak in detecting small targets while speeding up the convergence of the model. Finally, five more common types of defects were selected for training in the GuangDong TianChi fabric defect dataset, and using our proposed PEI-YOLOv5 with only 0.2 Giga Floating Point Operations (GFLOPs) increase, the mAP improved by 3.61%, reaching 87.89%. To demonstrate the versatility of PEI-YOLOv5, we additionally evaluated this in the NEU surface defect database, with the mAP of 79.37%. The performance of PEI-YOLOv 5 in these two datasets surpasses the most advanced fabric defect detection methods at present. We deployed the model to the NVIDIA Jetson TX2 embedded development board, and the detection speed reached 31 frames per second (Fps), which can fully meet the speed requirements of real-time detection.

Implementation of Convolutional Neural Network (CNN) for Image Classification of Leaf Disease In Mango Plants Using Deep Learning Approach

Plant diseases pose a serious threat to a country's economy and food security. One way to identify diseases in plants is through the visible features on their leaves. Farmers need to conduct an active examination of the condition of the leaves of plants to eradicate this disease. In this case, automatic recognition and classification of diseases of leaf crops is required in order to obtain an accurate identification. Digital image processing technology can be used to solve this problem. One effective approach is the Convolutional Neural Network (CNN). The trial image used a dataset consisting of 4000 images of mango leaf disease, namely Anthracnose, Bacterial Canker, Cutting Weevil, Die Back, Gall Midge, Powdery Mildew, and Sooty Mould. This study aims to compare the accuracy of CNN, VGG16 and InceptionV3. Architectural modeling uses these drawings to train and test models in recognizing and classifying mango leaf diseases. The results of modeling trials in the three scenarios were most optimally obtained by VGG16 with an accuracy of 96.87%, then InceptionV3 with an acquisition of 96.50% and CNN by 81%.

A lightweight convolutional neural network for automated crack inspection

Convolutional neural networks (CNN) have been increasingly embraced in building inspection and maintenance to realize automated crack inspection. However, CNNs are still less frequently utilized in the construction industry because most of them are designed with complex architectures that need to be executed by high-powered and expensive GPUs. To improve its practical applicability, this study aims to design a CNN model with a lightweight architecture for crack inspection. The lightweight design is expected to be feasible on a variety of electronic devices, such as mobile phones. To do so, layer pruning and parameter reduction methods were employed to design model depth and weight, respectively. Specifically, the CNN depth was first determined by training and testing 24 experimental groups that contain sequentially increased convolutional and pooling blocks. Secondly, 4 × 4 convolutional filters with 2 strides were inserted between adjacent convolutional layers to reduce model weight and memory. As a result, only 6.82 million weights and 24 M of memory were remined, compared to 138 million weights and 421 M of memory in the widely utilized model VGG-16. During an on-site validation, the proposed lightweight CNN was successfully executed on a mobile robot powered by a CPU-driven robotic control board. Both cracked and non-cracked surfaces could be detected automatically, remotely, and in real-time. Outstanding performance was also demonstrated, with F1-scores in the training and testing datasets of 96.8% and 92.4%, respectively. Overall, the designed lightweight CNN model is validated as more practicable for on-site inspection devices, which facilitate the wide spread of computer vision-based technologies in the construction industry. Meanwhile, the findings provide a potential solution for balancing the weight and accuracy of CNNs.

A CNN Hyperparameters Optimization Based on Particle Swarm Optimization for Mammography Breast Cancer Classification.

Breast cancer is considered one of the most-common types of cancers among females in the world, with a high mortality rate. Medical imaging is still one of the most-reliable tools to detect breast cancer. Unfortunately, manual image detection takes much time. This paper proposes a new deep learning method based on Convolutional Neural Networks (CNNs). Convolutional Neural Networks are widely used for image classification. However, the determination process for accurate hyperparameters and architectures is still a challenging task. In this work, a highly accurate CNN model to detect breast cancer by mammography was developed. The proposed method is based on the Particle Swarm Optimization (PSO) algorithm in order to look for suitable hyperparameters and the architecture for the CNN model. The CNN model using PSO achieved success rates of 98.23% and 97.98% on the DDSM and MIAS datasets, respectively. The experimental results proved that the proposed CNN model gave the best accuracy values in comparison with other studies in the field. As a result, CNN models for mammography classification can now be created automatically. The proposed method can be considered as a powerful technique for breast cancer prediction.

SHAP values accurately explain the difference in modeling accuracy of convolution neural network between soil full-spectrum and feature-spectrum

Acquiring soil nutrient content quickly and accurately through remote sensing is the key to advance precision agriculture. The development of deep learning has provided new technical means for soil hyperspectral modeling. However, the problem of poor interpretability of deep learning models limits its development. Although SHapley Additive exPlanations (SHAP) values based on game theory have been successfully applied to the interpretation of deep learning soil spectral modeling, whether they can accurately explain the differences in deep learning model accuracy remains to be verified. Based on this, we explored whether SHAP values can accurately explain the differences in convolutional neural network (CNN) modeling accuracy. We collected soil samples from agricultural land in the Liangshui River Basin in the southern mountainous and hilly areas of China, and measured the soil total nitrogen (STN) content and soil spectral data in the laboratory. We compared the effects of full-spectrum and feature-spectrum on the accuracy of deep learning models, and obtained the contribution of wavelengths in the CNN modeling process by calculating SHAP values. The results showed that combining different spectral pre-processing methods can play their respective advantages and help improve modeling accuracy. Among them, the CNN model obtained the highest prediction accuracy under the first-derivative Savitzky-Golay smoothing combination standard normal variate (SG1-SNV) spectral pre-processing in full-spectrum modeling. Compared with the feature-spectrum selected for modeling by Mutual information (MI) and competitive adaptive reweighted sampling (CARS), the CNN model achieved higher accuracy in most pre-processed spectra in full-spectrum modeling, and SHAP values accurately explained this reason. This is because the contribution is usually higher at most wavelengths with a high correlation with STN content. The feature-spectrum selected by CARS is more widely distributed but lacks continuity, and some wavelengths with high correlation and high contribution will also be missed. Meanwhile, some wavelengths with low correlation also have high contributions, which are usually not involved in the feature spectrum modeling of MI, thus affecting the modeling accuracy. Therefore, the deep learning model is more suitable for full-spectrum modeling due to its strong feature extraction and self-learning capabilities, and SHAP can obtain the wavelength contribution of the CNN model in soil spectral modeling, and then explain the differences in modeling accuracy. This study further proves the interpretability of deep learning, provides an important basis for the application of deep learning in soil hyperspectral modeling.