Single-channel Images Research Articles

WTDUN: Wavelet Tree-Structured Sampling and Deep Unfolding Network for Image Compressed Sensing

Deep unfolding networks have gained increasing attention in the field of compressed sensing (CS) owing to their theoretical interpretability and superior reconstruction performance. However, most existing deep unfolding methods often face the following issues: 1) they learn directly from single-channel images, leading to a simple feature representation that does not fully capture complex features; and 2) they treat various image components uniformly, ignoring the characteristics of different components. To address these issues, we propose a novel wavelet-domain deep unfolding framework named WTDUN, which operates directly on the multi-scale wavelet subbands. Our method utilizes the intrinsic sparsity and multi-scale structure of wavelet coefficients to achieve a tree-structured sampling and reconstruction, effectively capturing and highlighting the most important features within images. Specifically, the design of tree-structured reconstruction aims to capture the inter-dependencies among the multi-scale subbands, enabling the identification of both fine and coarse features, which can lead to a marked improvement in reconstruction quality. Furthermore, a wavelet domain adaptive sampling method is proposed to greatly improve the sampling capability, which is realized by assigning measurements to each wavelet subband based on its importance. Unlike pure deep learning methods that treat all components uniformly, our method introduces a targeted focus on important subbands, considering their energy and sparsity. This targeted strategy lets us capture key information more efficiently while discarding less important information, resulting in a more effective and detailed reconstruction. Extensive experimental results on various datasets validate the superior performance of our proposed method.

Three-stage binarization of color document images based on discrete wavelet transform and generative adversarial networks

The efficient extraction of text information from the background in degraded color document images is an important challenge in the preservation of ancient manuscripts. The imperfect preservation of ancient manuscripts has led to different types of degradation over time, such as page yellowing, staining, and ink bleeding, seriously affecting the results of document image binarization. This work proposes an effective three-stage network method to image enhancement and binarization of degraded documents using generative adversarial networks (GANs). Specifically, in Stage-1, we first split the input images into multiple patches, and then split these patches into four single-channel patch images (gray, red, green, and blue). Then, three single-channel patch images (red, green, and blue) are processed by the discrete wavelet transform (DWT) with normalization. In Stage-2, we use four independent generators to separately train GAN models based on the four channels on the processed patch images to extract color foreground information. Finally, in Stage-3, we train two independent GAN models on the outputs of Stage-2 and the resized original input images (512 × 512) as the local and global predictions to obtain the final outputs. The experimental results show that the Avg-Score metrics of the proposed method are 77.64, 77.95, 79.05, 76.38, 75.34, and 77.00 on the (H)-DIBCO 2011, 2013, 2014, 2016, 2017, and 2018 datasets, which are at the state-of-the-art level. The implementation code for this work is available at https://github.com/abcpp12383/ThreeStageBinarization.

A HRWS SAR Motion Compensation Method with Multichannel Phase Correction

The multichannel synthetic aperture radar (SAR) possesses the capability to acquire high-resolution, wide-swath SAR imagery, which has great potential for application. However, similar to traditional single-channel SAR systems, it suffers from imaging quality degradation due to motion errors. Many motion compensation algorithms have been used to improve the quality of single-channel SAR images, while fewer studies have been conducted on multichannel SAR motion compensation methods. The sub-image motion compensation method utilizes the single channel motion errors to perform multichannel motion errors compensation, considering that multiple channels have the same phase errors. To improve the quality of multichannel SAR imaging when multiple channel motion errors are inconsistent, this paper proposes a motion compensation method with multichannel phase correction for HRWS SAR. First, the method derives the phase errors estimation model via maximum sharpness to simultaneously estimate multichannel phase. Then, it compensates for the motion errors of all channels during backprojection imaging. The inconsistent motion errors of multiple channels can be compensated by estimating the phase errors of all channels, improving the image quality. The channel phase errors can be corrected while compensating for the motion errors. Moreover, the experimental results of point targets and complex scenes validate the effectiveness of the proposed method.

Color image hiding with invisible visual cryptography

In this paper, we propose a wavelength multiplexing image hiding system based on color invisible visual cryptography (CIVC). The image is divided into red, green, and blue channels. Each channel turns into several meaningful share images, different from the random, meaningless share images in visual cryptography (VC), and these are concealed within phase keys. The acquired single-channel share images, once synthesized into a color image and subjected to incoherent superposition, enable the extraction of the secret color image. Optical experiments are presented to demonstrate the feasibility and effectiveness of the proposed method.

Reconstruction of OFDM Signals Using a Dual Discriminator CGAN with BiLSTM and Transformer.

Communication signal reconstruction technology represents a critical area of research within communication countermeasures and signal processing. Considering traditional OFDM signal reconstruction methods' intricacy and suboptimal reconstruction performance, a dual discriminator CGAN model incorporating LSTM and Transformer is proposed. When reconstructing OFDM signals using the traditional CNN network, it becomes challenging to extract intricate temporal information. Therefore, the BiLSTM network is incorporated into the first discriminator to capture timing details of the IQ (In-phase and Quadrature-phase) sequence and constellation map information of the AP (Amplitude and Phase) sequence. Subsequently, following the addition of fixed position coding, these data are fed into the core network constructed based on the Transformer Encoder for further learning. Simultaneously, to capture the correlation between the two IQ signals, the VIT (Vision in Transformer) concept is incorporated into the second discriminator. The IQ sequence is treated as a single-channel two-dimensional image and segmented into pixel blocks containing IQ sequence through Conv2d. Fixed position coding is added and sent to the Transformer core network for learning. The generator network transforms input noise data into a dimensional space aligned with the IQ signal and embedding vector dimensions. It appends identical position encoding information to the IQ sequence before sending it to the Transformer network. The experimental results demonstrate that, under commonly utilized OFDM modulation formats such as BPSK, QPSK, and 16QAM, the time series waveform, constellation diagram, and spectral diagram exhibit high-quality reconstruction. Our algorithm achieves improved signal quality while managing complexity compared to other reconstruction methods.

DEM super-resolution guided by shaded relief using attention-based fusion

Deep-learning based approaches have been proven effective for Digital Elevation Model (DEM) super-resolution (SR) tasks. Previous networks typically treat DEM elevation values as single-channel image for input. However, DEM images alone cannot fully capture spatial and terrain features. Shaded relief images (SRIs), derived from DEMs, serve as crucial visual cues that intuitively convey terrain characteristics, addressing the limitations of DEM images and providing synergistic benefits for training DL models. The primary challenge in utilizing SRIs for guiding DEM SR lies in accurately selecting a consistent structure to extract and effectively integrate features from SRIs and DEMs. In this study, we propose an Attention-based Hierarchical Terrain Fusion (AHTF) framework for guided DEM SR. Specifically, an Attention-based Feature Fusion Module (AFFM) is designed to efficiently fuse relevant information from LR DEM and SRI, which includes a feature enhancement block to select valuable features and a feature recalibration block to fuse diverse terrain features. Additionally, we optimize the loss function from the perspectives of terrain analysis and visual effects. We validate AHTF on our newly constructed real-world Shade-DEM SR dataset and two open-source DEM SR datasets. Compared to the current state-of-the-art methods, our AHTF achieves the best results in terms of root mean square error (RMSE) for elevation, slope, and aspect. Furthermore, the extracted stream networks are closer to real-world conditions. This study offers new insights and methods for further research and application in the field of DEM super-resolution. Our dataset can be obtained at https://doi.org/10.6084/m9.figshare.25590945.

Brain-computer interfaces inspired spiking neural network model for depression stage identification

BackgroundDepression is a global mental disorder, and traditional diagnostic methods mainly rely on scales and subjective evaluations by doctors, which cannot effectively identify symptoms and even carry the risk of misdiagnosis. Brain-Computer Interfaces inspired deep learning-assisted diagnosis based on physiological signals holds promise for improving traditional methods lacking physiological basis and leads next generation neuro-technologies. However, traditional deep learning methods rely on immense computational power and mostly involve end-to-end network learning. These learning methods also lack physiological interpretability, limiting their clinical application in assisted diagnosis. MethodologyA brain-like learning model for diagnosing depression using electroencephalogram (EEG) is proposed. The study collects EEG data using 128-channel electrodes, producing a 128×128 brain adjacency matrix. Given the assumption of undirected connectivity, the upper half of the 128×128 matrix is chosen in order to minimise the input parameter size, producing 8,128-dimensional data. After eliminating 28 components derived from irrelevant or reference electrodes, a 90×90 matrix is produced, which can be used as an input for a single-channel brain-computer interface image. ResultAt the functional level, a spiking neural network is constructed to classify individuals with depression and healthy individuals, achieving an accuracy exceeding 97.5 %. Comparison with existing methodsCompared to deep convolutional methods, the spiking method reduces energy consumption. ConclusionAt the structural level, complex networks are utilized to establish spatial topology of brain connections and analyse their graph features, identifying potential abnormal brain functional connections in individuals with depression.

Multi-channel and multi-featured extended orb with gabor filter and non-maximum suppression

This research paper presents an enhanced approach to feature extraction using the Oriented FAST and Rotated BRIEF (ORB) algorithm (Rublee et al., 2014). The proposed method leverages the power of Gabor filters (Mehrotra et al., 1992) in combination with Non-Maximum Suppression (Neuback and Van Gool, 2006) to improve the robustness and efficiency of feature detection in computer vision applications. The process begins with the acquisition of an RGB image, which is then transformed into seven single-channel images representing different color and intensity aspects: red, green, blue, hue, saturation, value, and grayscale. Each single-channel image is independently subjected to Gabor filtering, resulting in seven Gabor-filtered images. These images capture distinct texture and frequency information, enhancing the discriminative power of subsequent feature extraction. Subsequently, the Oriented FAST and Rotated BRIEF (ORB) algorithm is applied to each of the seven Gabor-filtered images to extract key points and descriptors. This step yields a comprehensive set of keypoints (Mallick et al., 2015) and descriptors, capturing rich local feature information across multiple image channels. To further refine the extracted features and eliminate redundant keypoints, Non-Maximum Suppression is employed independently on each single-channel image. This process effectively filters out overlapped and false keypoints, ensuring the selection of only the most salient features for downstream tasks. Experimental results demonstrate the efficacy of the proposed approach in improving feature detection performance compared to traditional methods. The combination of Gabor filtering and Non-Maximum Suppression enhances feature discriminability, robustness to noise, and resistance to image transformations, making it well-suited for various computer vision applications, including object recognition, image matching, and scene understanding.

Deep-learning-based motion correction using multichannel MRI data: a study using simulated artifacts in the fastMRI dataset.

Deep learning presents a generalizable solution for motion correction requiring no pulse sequence modifications or additional hardware, but previous networks have all been applied to coil-combined data. Multichannel MRI data provide a degree of spatial encoding that may be useful for motion correction. We hypothesize that incorporating deep learning for motion correction prior to coil combination will improve results. A conditional generative adversarial network was trained using simulated rigid motion artifacts in brain images acquired at multiple sites with multiple contrasts (not limited to healthy subjects). We compared the performance of deep-learning-based motion correction on individual channel images (single-channel model) with that performed after coil combination (channel-combined model). We also investigate simultaneous motion correction of all channel data from an image volume (multichannel model). The single-channel model significantly (p < 0.0001) improved mean absolute error, with an average 50.9% improvement compared with the uncorrected images. This was significantly (p < 0.0001) better than the 36.3% improvement achieved by the channel-combined model (conventional approach). The multichannel model provided no significant improvement in quantitative measures of image quality compared with the uncorrected images. Results were independent of the presence of pathology, and generalizable to a new center unseen during training. Performing motion correction on single-channel images prior to coil combination provided an improvement in performance compared with conventional deep-learning-based motion correction. Improved deep learning methods for retrospective correction of motion-affected MR images could reduce the need for repeat scans if applied in a clinical setting.

Contrast enhancement algorithm for infrared images based on atmospheric scattering model

With the rapid development of infrared technology, infrared cameras are widely used in the military, medical, and industrial fields. However, the thermal radiation information received by infrared cameras often undergoes degradation due to random factors such as scattering or reflection during transmission. Consequently, the final infrared image suffers from poor contrast and blurred vision. To address these problems, this paper proposes an infrared image contrast enhancement algorithm based on the atmospheric scattering model. Firstly, we modify the hypothesis of the dark channel prior theory, and the pseudo dark channel prior theory is deduced to meet the processing requirements of single-channel images. Secondly, we propose anti-target interference and filtering threshold method to accurately estimate the atmospheric illumination value and transmittance of the atmospheric scattering model. These methods effectively mitigate the influence of infrared targets on the atmospheric illumination value and eliminate the blocking artifact effect caused by pixel value jumps in the transmittance images. Furthermore, we employ adaptive histogram equalization with contrast limitation, highlighting the global contrast. This paper conducts experiments on the M3FD dataset and employs objective evaluations using Visible Edge (E), Figure Definition (FD), Peak Signal-to-Noise Ratio (PSNR), Information Entropy (IE), Structure Similarity Index Measure (SSIM), and Visual Information Fidelity (VIF). The results indicate that the proposed algorithm enhances the overall contrast of images while highlighting fine texture details, exhibiting good visual effects. Both subjective and objective evaluations outperform seven other contrast algorithms, demonstrating the effectiveness of the proposed method.

Enhancing subcellular protein localization mapping analysis using Sc2promap utilizing attention mechanisms

BackgroundAberrant protein localization is a prominent feature in many human diseases and can have detrimental effects on the function of specific tissues and organs. High-throughput technologies, which continue to advance with iterations of automated equipment and the development of bioinformatics, enable the acquisition of large-scale data that are more pattern-rich, allowing for the use of a wider range of methods to extract useful patterns and knowledge from them. MethodsThe proposed sc2promap (Spatial and Channel for SubCellular Protein Localization Mapping) model, designed to proficiently extract meaningful features from a vast repository of single-channel grayscale protein images for the purposes of protein localization analysis and clustering. Sc2promap incorporates a prediction head component enriched with supplementary protein annotations, along with the integration of a spatial-channel attention mechanism within the encoder to enables the generation of high-resolution protein localization maps that encapsulate the fundamental characteristics of cells, including elemental cellular localizations such as nuclear and non-nuclear domains. ResultsQualitative and quantitative comparisons were conducted across internal and external clustering evaluation metrics, as well as various facets of the clustering results. The study also explored different components of the model. The research outcomes conclusively indicate that, in comparison to previous methods, Sc2promap exhibits superior performance. ConclusionsThe amalgamation of the attention mechanism and prediction head components has led the model to excel in protein localization clustering and analysis tasks. General significanceThe model effectively enhances the capability to extract features and knowledge from protein fluorescence images.

A series fault arc detection method based on denoising autoencoder and deep residual network

Given the problem that the existing series arc fault identification methods use existing features such as the time-frequency domain of the current signal as the basis for identification, resulting in relatively limited arc detection solutions, and that the methods of directly extracting current signal features using deep learning algorithms have insufficient feature extraction, a new series arc fault detection method based on denoising autoencoder (DAE) and deep residual network (ResNet) is proposed. First, a large number of training samples are obtained through sliding window and data normalization methods, and then high-dimensional abstract feature data are obtained from the fault and normal samples collected in the experiment through denoising autoencoders, converted into grayscale images, and processed in pseudo-color. The single-channel grayscale images are mapped into three-channel color values, and finally, the three-channel values are input into the constructed deep residual network for deep learning training. In the 152 super high-level ResNet, the arc fault recognition rate can reach 99.7%. For loads that have not participated in ResNet network training, the recognition rate can also reach 97.6%.

Supervised learning study on ground classification and state recognition of agricultural robots based on multi-source vibration data fusion

In agricultural environments, recognizing the walking ground and state of tracked mobile robots is a complex and challenging task, influenced by clay conditions and other external environmental disturbances. Therefore, this paper proposes a novel data processing method and efficient classifier. Firstly, the noise signals on the left and right sides of the collected robot are averaged, and a time-wavelet-time domain transformation is performed using the Mallat algorithm to achieve nonlinear enhancement of data features and signal denoising. Secondly, the Gramian Angular Summation Fields (GASF) is introduced to transform sequence data into single-channel images, capturing the periodicity and similarity of time series. Next, the images of three sets of sequences are stacked in the channel dimension in RGB format, thus achieving feature fusion of multi-source data. Finally, a supervised learning classifier named Attention-fused Residual Convolutional Neural Network (ANR-CNN) is proposed. Here, the combination of channel and spatial attention mechanisms captures important features in the feature map in both channel and spatial dimensions. The convolutional residual structure enhances feature transmission, improving the model's classification accuracy. Experimental results demonstrate that the proposed data augmentation method effectively enhances model performance, and the classification accuracy of ANR-CNN reaches 92.35%. This implies accurate recognition of the walking ground and state of tracked mobile robots in agricultural environments.

A multi-tasking novel variational model for image decolorization and denoising

Image decolorization has a wide range of applications in digital printing, photo rendering and single-channel image processing. During the process of image decolorization, it is significant to retain more contrast information and avoid increasing the noise level of the original image. As we know, noise was hardly considered in previous decolorization methods. In this paper, we propose a new multi-tasking variational model with two tasks: decolorizing and denoising. The first task is image decolorization, that is, to transform color images to grayscale images. The other task is to control the noise level, and even eliminate the noise. Inspired by TV-Stokes model, we introduce the unit normal vector containing the basic geometry information of the color image in the model. The existence and the uniqueness of the solution to the variational problems are proved. For the variational model, two numerical methods are adopted. At first, we use gradient descent method to obtain evolution equations and construct finite difference schemes to solve the corresponding equations of proposed variational model. Furthermore, an efficient augmented Lagrangian approach is applied to solve the multi-tasking variational model directly. For noisy images and clean images, we employ two output patterns of grayscale images respectively. Especially, our model is stable for the noisy images. Numerical results and comparisons show that the proposed model can produce highly competitive results in terms of decolorization quality of noisy images, edge preservation and adjustment of parameters.

Detection of early bruises on apples using near‐infrared camera imaging technology combined with adaptive threshold segmentation algorithm

AbstractApples are highly susceptible to bruising during picking, packing, and transportation, and early detection of apple bruises allows timely screening and treatment to reduce food safety risks. However, early bruises on apples, especially bruises of 30 min or less, with similar appearance to sound tissue, making manual sorting difficult. To address this problem, a near‐infrared (NIR) camera imaging technology combined with adaptive threshold segmentation algorithm for early apple bruise detection is proposed in this article. First, a NIR camera imaging system is built to acquire full‐band images of fruit face, fruit stem face and calyx face of sound apples and apples with single bruise and multiple bruises at 30 min of occurrence. Then, an adaptive threshold segmentation algorithm is proposed and an image processing algorithm, which includes high‐pass filtering, feature extraction, adaptive threshold segmentation and Hough circle detection, is developed and applied to full‐band images in response to the interference of apple stem and calyx on early bruising detection in this article. The results show that the proposed algorithm has an accuracy of 95.56% for the detection of sound apple samples and an F1‐score of 94.70% for the detection of early bruised samples, which is better than the commonly used image segmentation algorithms such as Otsu and fixed thresholding. This study shows that the combination of NIR camera imaging and the adaptive threshold segmentation algorithm proposed in this article can effectively detect early bruises on apples, providing a new idea for rapid and nondestructive detection of agricultural product quality.Practical applicationsDetecting early bruises on apples can achieve effective monitoring and control of apple supply chain, and guarantee the quality and safety of agricultural products. Most studies are based on conventional machine vision with RGB camera or hyperspectral imaging technology, while conventional machine vision based on RGB camera is difficult to detect apple early bruises that are not visible to the naked eye, and hyperspectral imaging has a large amount of data and redundant image data among wavelengths. Compared with previous studies, apple early bruise detection method proposed in this article combining NIR camera imaging technology and adaptive threshold segmentation algorithm can avoid the interference of bright spots, fruit stem and calyx on bruise segmentation while capturing a single channel image, providing a new method for accurate and rapid detection of early bruises on apples.

Optical single-channel color image encryption based on chaotic palmprint phase masks

Optical single-channel color image encryption based on chaotic palmprint phase masks

3D Human Pose Estimation from multi-view thermal vision sensors

Human Pose Estimation from images allows the recognition of key daily activity patterns in Smart Environments. Current State-of-the-art (SOTA) 3D pose estimators are built on visible spectrum images, which can lead to privacy concerns in Ambient Assisted Living solutions. Thermal Vision sensors are being deployed in these environments, as they preserve privacy and operate in low brightness conditions. Furthermore, multi-view setups provide the most accurate 3D pose estimation, as the occlusion problem is overcome by having images from different perspectives. Nevertheless, no solutions in the literature use thermal vision sensors following a multi-view scheme.In this work, a multi-view setup consisting of low-cost devices is deployed in the Smart Home of the University of Almería. Thermal and visible images are paired using homography, and SOTA solutions such as YOLOv3 and Blazepose are used to annotate the bounding box and 2D pose in the thermal images. ThermalYOLO is built by fine-tuning YOLOv3 and outperforms YOLOv3 by 5% in bounding box recognition and by 1% in IoU value. Furthermore, InceptionResNetV2 is found as the most appropriate architecture for 2D pose estimation.Finally, a 3D pose estimator was built comparing input approaches and convolutional architectures. Results show that the most appropriate architecture is having three single-channel thermal images processed by independent convolutional backbones (ResNet50 in this case). After these, the output is fused with the 2D poses. The resulting convolutional neural network shows excellent behaviour when having occlusions, outperforming single-view SOTA approaches in the visible spectrum.

A review of recent advance of ship detection in single-channel SAR images

Synthetic aperture radar (SAR) is an active microwave imaging sensor for high-resolution observation, with the ability of working in all-weather and all-day. Recently, SAR images have been widely used in many fields. Among them, ship detection in single-channel SAR images is a significant part of civilian and military fields. This article first discusses the characteristic of SAR images and the detectability of ships, then summarizes the recent advance of traditional and deep learning-based methods used for ship detection in single-channel SAR images. In addition, the characteristics and existing problems of various methods are discussed and their future development trends are predicted. Aiming at the problems of the large amount of calculation, multi-scale and densely docked ship detection in single-channel SAR images, an improved deep learning-based detection algorithm is proposed, which has achieved excellent performance on the SAR ship detection dataset (SSDD).

Detection and segmentation of radio frequency interference from satellite images using attention-GANs

Radio frequency interference (RFI) refers to the interference in the electromagnetic spectrum caused by undesired radio signals that share the same frequency band as the desired signal. RFI can interrupt, distort, or completely eliminate radio signals, resulting in ineffective communication or interaction with other electronic systems. RFI can have several effects on satellite imagery. It can result in stripes, smears, and other visual patterns, degrading image quality and making it harder to understand. In extreme cases, RFI can entirely obscure visual characteristics of interest. Among the different techniques for identifying RFIs viz thresholding, detection using artificial intelligence (AI) has shown higher efficiency and accuracy with minimal manual intervention. This paper proposes an RFI detection and segmentation model based on an attention-enforced Generative Adversarial Network (GAN): Attention-SegmentationGAN(SegGAN), using an image-to-image translation network: Pix2Pix. This approach turns the RFI segmentation into an image translation issue and then trains two deep convolutional neural networks (CNN), concurrently to provide a binary RFI mask image. Furthermore, we enhance the network architectures of the Pix2Pix model’s generator and discriminator for superior quality detection and localization of RFI from single-channel satellite images. Here, a dataset consisting of around 1000 single-channel satellite images with synthetically generated RFI in random portions of the image was used. For model developments, Standard metrics such as Sensitivity(Recall), Specificity(Accuracy), Precision, and Dice coefficient have been utilized to evaluate the detection of noise-affected locations. The proposed model showed superior segmentation performance with the baseline Pix2Pix model and state-of-the-art RFI detection approaches from satellite images.

Gray-to-color image conversion in the classification of breast lesions on ultrasound using pre-trained deep neural networks.

Breast ultrasound (BUS) image classification in benign and malignant classes is often based on pre-trained convolutional neural networks (CNNs) to cope with small-sized training data. Nevertheless, BUS images are single-channel gray-level images, whereas pre-trained CNNs learned from color images with red, green, and blue (RGB) components. Thus, a gray-to-color conversion method is applied to fit the BUS image to the CNN's input layer size. This paper evaluates 13 gray-to-color conversion methods proposed in the literature that follow three strategies: replicating the gray-level image to all RGB channels, decomposing the image to enhance inherent information like the lesion's texture and morphology, and learning a matching layer. Besides, we introduce an image decomposition method based on the lesion's structural information to describe its inner and outer complexity. These gray-to-color conversion methods are evaluated under the same experimental framework using a pre-trained CNN architecture named ResNet-18 and a BUS dataset with more than 3000 images. In addition, the Matthews correlation coefficient (MCC), sensitivity (SEN), and specificity (SPE) measure the classification performance. The experimental results show that decomposition methods outperform replication and learning-based methods when using information from the lesion's binary mask (obtained from a segmentation method), reaching an MCC value greater than0.70 and specificity up to0.92, although the sensitivity is about0.80. On the other hand, regarding the proposed method, the trade-off between sensitivity and specificity is better balanced, obtaining about0.88 for both indices and an MCC of0.73. This study contributes to the objective assessment of different gray-to-color conversion approaches in classifying breast lesions, revealing that mask-based decomposition methods improve classification performance. Besides, the proposed method based on structural information improves the sensitivity, obtaining more reliable classification results on malignant cases and potentially benefiting clinical practice.