3. Producción
Browse
6 results
Search Results
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Application of semantic segmentation with few labels in the detection of water bodies from PeruSat-1 satellite’s images(Copernicus Publications, 2020-11-06)Abstract. Remote sensing is widely used to monitor earth surfaces with the main objective of extracting information from it. Such is the case of water surface, which is one of the most affected extensions when flood events occur, and its monitoring helps in the analysis of detecting such affected areas, considering that adequately defining water surfaces is one of the biggest problems that Peruvian authorities are concerned with. In this regard, semiautomatic mapping methods improve this monitoring, but this process remains a time-consuming task and into the subjectivity of the experts.In this work, we present a new approach for segmenting water surfaces from satellite images based on the application of convolutional neural networks. First, we explore the application of a U-Net model and then a transfer knowledge-based model. Our results show that both approaches are comparable when trained using an 680-labelled satellite image dataset; however, as the number of training samples is reduced, the performance of the transfer knowledge-based model, which combines high and very high image resolution characteristics, is improved. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Multi-view data augmentation to improve wound segmentation on 3D surface model by deep learning(Institute of Electrical and Electronics Engineers Inc., 2021-01-01)Wound area segmentation really progressed with the emergence of deep learning, due to its robustness in uncontrolled lighting and no need to design hand-crafted features but two limits have still to be overcome: firstly, its performance relies on the size and quality of the training dataset in the medical field, where data annotation is costly and time-consuming; secondly the accuracy of the segmentation depends highly on the camera distance and angle and moreover perspective effects prevent measuring real surfaces in single views. To address concurrently these two issues, we propose to apply multi-view modeling: an image sequence is acquired around the wound site and enables wound 3D reconstruction. Then, a segmentation step is run to extract roughly the wound from the background in each view and to select the best view with an original strategy. This view provides the most accurate segmentation and the real wound bed area even on non planar wounds. Finally, this segmentation is backprojected in each view to generate a complete set of well annotated real images to reinforce the learning step of the neural network. In our experiments, we compare several strategies to select the best view in the image sequence. The proposed method, tested on a dataset of 270 images, outperforms standard deep learning approach based on a single view, as recorded with DICE index and IoU score which rise respectively from 36.53% to 86.3% and 29.48% to 77.09% for the wound class to achieve an overall DICE and IoU score of 93.04% and 86.61% including background class. These results attest to the robustness of our method and its improved accuracy in the wound segmentation task. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Building semantic segmentation using UNet convolutional network on SpaceNet public data sets for monitoring surrounding area of Chan Chan (Peru)(AGH University of Science and Technology Press, 2024-01-01)The amount of damage to cultural heritage sites is increasing rapidly every year. This is due to inadequate heritage management and uncontrolled urban growth as well as unpredictable seismic and atmospheric events that manifest themselves in a continuously deteriorating ecosystem. Thus, applications of artificial intelligence (AI) in remote-sensing (RS) techniques (machine-learning and deep-learning algorithms) for monitoring archaeological sites have increased in recent years. This research involves the surrounding area of the archaeological site of Chan Chan in Peru in particular. An approach that is based on the use of AI algorithms for building footprint segmentation and change-detection analysis by means of RS images is proposed. It involves a UNet con-volutional network based on an EfficientNet B0 to B7 encoder. The network was trained on two public data sets from SpaceNet that were based on WV2 and WV3 satellite images: SpaceNet V1 (Rio), and SpaceNet V2 (Shanghai). In the pre-processing phase, the images from the two data sets have been equalized in order to improve their quality and avoid overfitting. The building segmentation has been performed on HRV images of the study area that were downloaded from Google Earth Pro. The value that was achieved in the IoU metric was around 70% in both experiments. The purpose of this proposed methodology is to assist scientists in drafting monitoring and conservation protocols based on already-recorded data in order to prevent future disasters and hazards. - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Morphology-enhanced CAM-guided SAM for weakly supervised breast lesion segmentation(Elsevier BV, 2026-05-01)Ultrasound imaging is vital for the early detection of breast cancer, where accurate lesion segmentation supports clinical diagnosis and treatment planning. However, existing deep learning-based methods rely on pixel-level annotations, which are costly and labor-intensive to obtain. This study presents a weakly supervised framework for breast lesion segmentation in ultrasound images. The framework combines morphological enhancement with Class Activation Map (CAM)-guided lesion localization and utilizes the Segment Anything Model (SAM) for refined segmentation without pixel-level labels. By adopting a lightweight region synthesis strategy and relying solely on SAM inference, the proposed approach substantially reduces model complexity and computational cost while maintaining high segmentation accuracy. Experimental results on the BUSI dataset show that our method achieves a Dice coefficient of 0.7063 under five-fold cross-validation and outperforms several fully supervised models in Hausdorff distance metrics. These results demonstrate that the proposed framework effectively balances segmentation accuracy, computational efficiency, and annotation cost, offering a practical and low-complexity solution for breast ultrasound analysis. The code for this study is available at: https://github.com/YueXin18/MorSeg-CAM-SAM-Segmentation.1 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Cost-effective and portable device for partial shading assessment in photovoltaic modules using RGB imaging(Elsevier BV, 2026-02-01)Partial shading conditions in photovoltaic systems can cause major loss in performance. Therefore, models to estimate the reduction and distribution of irradiance due to partial shading are important to evaluate and minimize losses. In this paper, a model to estimate the irradiance reduction under partial shading conditions on PV installations is discussed. The proposed approach utilizes RGB images in the visible spectrum to estimate irradiance within the shaded areas of individual modules. The results suggest that there is a significant correlation between gray scale levels and irradiance in the shaded area. The conclusion is based on comprehensive observations and data collected during the experimental phase, which demonstrate regularity in irradiance values within the shaded area. The key advantage of the proposed methodology is its practical framework, which enables it to address various problems and types of partial shading, regardless of the complexity of the geometry. Furthermore, a model was proposed and validated based on this methodology. It is demonstrated that considering partial shading through RGB images, and by employing segmentation, homogenization, and regression using analytic models, this leads to an estimation accuracy in the irradiance parameter (RMSE of ∼1 %) of 11.14 W/m2. This innovative methodology eliminates the need for multiple irradiance sensors or an IV curve tracer to map shadowed zones.3 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, An AI-enabled comprehensive breast ultrasound diagnostic system for low-resource settings without a sonographer or a radiologist(Nature Portfolio, 2026-12-01)Breast cancer is the most common non-skin related malignancy and the leading cause of cancer death in women. Mammography remains the gold standard for early detection; however, its accessibility is limited in low-resource settings due to cost and technical complexity. Ultrasound (US) is a viable alternative, but its implementation is hindered by the scarcity of trained radiologists and sonographers. Volume sweep imaging (VSI) has addressed the issue of US acquisition by enabling non-specialists to perform standardized scans. However, these still require expert interpretation, limiting their impact. To overcome this barrier, we propose a fully automated Breast VSI (VSI-B) system integrating artificial intelligence (AI) for segmentation and classification of breast lesions, aiming to provide an accessible diagnostic tool for low-resource environments. This study developed an AI-driven diagnostic system for VSI-B, combining a segmentation model (Attention U-Net 3D) with a classification model for lesion detection. A total of 98 patients with palpable breast lumps were included in the study. The dataset consisted of 392 VSI-B US videos and 2,100 classified frames. A new method was implemented to enhance mass identification by selecting key frames for analysis. A majority voting algorithm was used to optimize lesion classification. The system's performance was assessed based on sensitivity, specificity, and accuracy. Following a detection step that achieved 100% sensitivity and 93.6% specificity for cancer and no cancer patients, as well as 95.0% sensitivity and 63.0% specificity for mass and no mass patients, the DenseNet classification model reached 87% accuracy, 100% sensitivity, and 83% specificity. A majority voting algorithm optimized classification, yielding an AUC of 0.91. This study highlights the potential of an AI-enabled VSI-B system as a reliable diagnostic tool in low-resource settings. By integrating multi-modal segmentation and classification, the system automates breast lesion detection and stratification, reducing reliance on radiologists. The results suggest that this approach could enhance early breast cancer diagnosis and guide clinical decision-making, particularly in underserved regions.1