Beatriz Paulina Garcia-Salgado

Dr. Beatriz Paulina Garcia-Salgado

Instituto Politécnico Nacional - ESIME Culhuacán
+1
orcid

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
24 July 2021 in Remote Sensing
Verified
Journal Article
Open access
PDF file
CCClara Cruz-Ramos
Clara Cruz-Ramos
Beatriz Paulina Garcia-SalgadoBeatriz Paulina Garcia-Salgado
Beatriz Paulina Garcia-Salgado
RRRogelio Reyes-Reyes
Rogelio Reyes-Reyes
Volodymyr PonomaryovVolodymyr Ponomaryov
Volodymyr Ponomaryov
SSSergiy Sadovnychiy
Sergiy Sadovnychiy
The principles of the transform stage of the extract, transform and load (ETL) process can be applied to index the data in functional structures for the decision-making inherent in an urban remote sensing application. This work proposes a method that can be utilised as an organisation stage by reducing the data dimension with Gabor texture features extracted from grey-scale representations of the Hue, Saturation and Value (HSV) colour space and the Normalised Difference Vegetation Index (NDVI). Additionally, the texture features are reduced using the Linear Discriminant Analysis (LDA) method. Afterwards, an Artificial Neural Network (ANN) is employed to classify the data and build a tick data matrix indexed by the belonging class of the observations, which could be retrieved for further analysis according to the class selected to explore. The proposed method is compared in terms of classification rates, reduction efficiency and training time against the utilisation of other grey-scale representations and classifiers. This method compresses up to 87% of the original features and achieves similar classification results to non-reduced features but at a higher training time.

11 September 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
Beatriz Paulina Garcia-SalgadoBeatriz Paulina Garcia-Salgado
Beatriz Paulina Garcia-Salgado
JAJose Agustin Almaraz-Damian
Jose Agustin Almaraz-Damian
OCOscar Cervantes-Chavarria
Oscar Cervantes-Chavarria
Volodymyr PonomaryovVolodymyr Ponomaryov
Volodymyr Ponomaryov
RRRogelio Reyes-Reyes
Rogelio Reyes-Reyes
SSSergiy Sadovnychiy
Sergiy Sadovnychiy
Ischemic stroke lesion segmentation in MRI images represents significant challenges, particularly due to class imbalance between foreground and background pixels. Several approaches have been developed to achieve higher F1-Scores in stroke lesion segmentation under this challenge. These strategies include convolutional neural networks (CNN) and models that represent a large number of parameters, which can only be trained on specialized computational architectures that are explicitly oriented to data processing. This paper proposes a lightweight model based on the U-Net architecture that handles an attention module and the Generalized Dice Focal loss function to enhance the segmentation accuracy in the class imbalance environment, characteristic of stroke lesions in MRI images. This study also analyzes the segmentation performance according to the pixel size of stroke lesions, giving insights into the loss function behavior using the public ISLES 2015 and ISLES 2022 MRI datasets. The proposed model can effectively segment small stroke lesions with F1-Scores over 0.7, particularly in FLAIR, DWI, and T2 sequences. Furthermore, the model shows reasonable convergence with their 7.9 million parameters at 200 epochs, making it suitable for practical implementation on mid and high-end general-purpose graphic processing units.

2 April 2021 in Applied Sciences
Verified
Journal Article
Open access
PDF file
RRRogelio Reyes-Reyes
Rogelio Reyes-Reyes
CCClara Cruz-Ramos
Clara Cruz-Ramos
Volodymyr PonomaryovVolodymyr Ponomaryov
Volodymyr Ponomaryov
Beatriz Paulina Garcia-SalgadoBeatriz Paulina Garcia-Salgado
Beatriz Paulina Garcia-Salgado
JMJavier Molina-Garcia
Javier Molina-Garcia
In this paper, a fragile watermarking scheme for color image authentication and self-recovery with high tampering rates is proposed. The original image is sub-sampled and divided into non-overlapping blocks, where a watermark used for recovery purposes is generated for each one of them. Additionally, for each recovery watermark, the bitwise exclusive OR (XOR) operation is applied to obtain a single bit for the block authentication procedure. The embedding and extraction process can be implemented in three variants (1-LSB, 2-LSB or 3-LSB) to solve the tampering coincidence problem (TCP). Three, six or nine copies of the generated watermarks can be embedded according to the variant process. Additionally, the embedding stage is implemented in a bit adjustment phase, increasing the watermarked image quality. A particular procedure is applied during a post-processing step to detect the regions affected by the TCP in each recovery watermark, where a single faithful image used for recovery is generated. In addition, we involve an inpainting algorithm to fill the blocks that have been tampered with, significantly increasing the recovery image quality. Simulation results show that the proposed framework demonstrates higher quality for the watermarked images and an efficient ability to reconstruct tampered image regions with extremely high rates (up to 90%). The novel self-recovery scheme has confirmed superior performance in reconstructing altered image regions in terms of objective criteria values and subjective visual perception via the human visual system against other state-of-the-art approaches.

30 March 2020 in Journal of Applied Remote Sensing
Verified
Journal Article
Closed access
PDF file
Beatriz Paulina Garcia-SalgadoBeatriz Paulina Garcia-Salgado
Beatriz Paulina Garcia-Salgado
VPVolodymyr I. Ponomaryov
Volodymyr I. Ponomaryov
SSSergiy Sadovnychiy
Sergiy Sadovnychiy
RRRogelio Reyes-Reyes
Rogelio Reyes-Reyes
<!-- *** Custom HTML *** --> A large amount of remote sensing data can be easily acquired due to the increase in the advances in sensor’s technologies. The sensors can generate high-dimensional data in a lower time producing problems related to big data such as management and organization. Since the acquired data is characterized by a large dimension and lack of structure, the information analysis becomes harder. Therefore, an organization stage should structure the data reducing the dimension while maintaining the main properties to enable further analysis. The feature extraction and selection methods can achieve this task. Consequently, we aim to explore various pixel-wise feature extraction and selection algorithms to manage the organization stage of big data for hyperspectral images. Our work covers the comparison between feature vectors computed using the discrete Fourier transform, discrete cosine transform (DCT), and stationary wavelet transform. Moreover, spectral angle mapper, Jeffries–Matusita distance, spectral information divergence, and linear discriminant analysis (LDA) were implemented as feature selectors. Feature extraction and selection methods were combined and evaluated in terms of algorithm complexity, reduction efficiency, and classification accuracy with the aid of a support vector machine and a maximum likelihood classifier. The analysis shows that some linear transformations can perform better in natural landscapes and others in urban images. Furthermore, the study found that the combination of DCT and LDA, which achieves high classification rates with an efficient dimension reduction, can be suitable for the organization stage of a big data remote sensing application of hyperspectral images.

1 February 2016
Verified
Conference Paper
Closed access
Beatriz Paulina Garcia-SalgadoBeatriz Paulina Garcia-Salgado
Beatriz Paulina Garcia-Salgado
Volodymyr PonomaryovVolodymyr Ponomaryov
Volodymyr Ponomaryov
Hyperspectral images can be represented as a cube data structure. As a consequence, a spatial classification could be a difficult task. In this work, we describe a novel feature extraction methodology in order to perform a Hyperspectral image spatial classification. We turn the hyperspectral data into a gray-scaled HSV image and a Normalize Difference Vegetation Index (NDVI) representation. Afterwards, Haralick texture features are computed for both images, and the resulted features vectors are fused calculating the determinants of the matrices composed of these characteristics. To test the experimental accuracy of the proposed method, we employ five Hyperspectral images and a Maximum Likelihood Classifier (MLC). The current proposal is compared against other state-of-the-art methods, such as the employment of Principal Components Analysis (PCA).

7 July 2022 in Sensors
Verified
Journal Article
Open access
PDF file
RRRogelio Reyes-Reyes
Rogelio Reyes-Reyes
GAGibran Humberto Aranda-Bojorges
Gibran Humberto Aranda-Bojorges
Beatriz Paulina Garcia-SalgadoBeatriz Paulina Garcia-Salgado
Beatriz Paulina Garcia-Salgado
Volodymyr PonomaryovVolodymyr Ponomaryov
Volodymyr Ponomaryov
CCClara Cruz-Ramos
Clara Cruz-Ramos
SSSergiy Sadovnychiy
Sergiy Sadovnychiy
This work proposes a novel scheme for speckle suppression on medical images acquired by ultrasound sensors. The proposed method is based on the block matching procedure by using mutual information as a similarity measure in grouping patches in a clustered area, originating a new despeckling method that integrates the statistical properties of an image and its texture for creating 3D groups in the BM3D scheme. For this purpose, the segmentation of ultrasound images is carried out considering superpixels and a variation of the local binary patterns algorithm to improve the performance of the block matching procedure. The 3D groups are modeled in terms of grouped tensors and despekled with singular value decomposition. Moreover, a variant of the bilateral filter is used as a post-processing step to recover and enhance edges’ quality. Experimental results have demonstrated that the designed framework guarantees a good despeckling performance in ultrasound images according to the objective quality criteria commonly used in literature and via visual perception.