arXiv subjects

Manuel Pérez-Carrasco

Publications and source records attributed to Manuel Pérez-Carrasco.

9 recordsLinked to original sources

arXiv2026

Plume Segmentation from MethaneSAT with Cross-Sensor Transfer Learning and Physics-Informed Postprocessing

Automated detection and masking of individual methane plumes from satellite imagery is important for operational emission attribution and quantification. We present a machine learning framework for plume detection from MethaneSAT retrieved column-averaged dry-air mole fractions of methane. We address two core challenges: the scarcity of labeled MethaneSAT data and the need for inference reliability across diverse atmospheric and surface conditions. We first demonstrate that Mask R-CNN with a ResNet-50 backbone outperforms a U-Net semantic segmentation baseline on both MethaneAIR (an airborne version of MethaneSAT) and MethaneSAT data, with pixel-level F1 score gains of 10.49 and 5.48 percentage points respectively. To address MethaneSAT data scarcity, we evaluate three cross-sensor transfer strategies leveraging MethaneAIR flights and synthetic plumes. Mask R-CNN with ResNet-50 fine-tuned from MethaneAIR pre-trained weights is the most effective strategy, achieving instance-level precision of 0.60 and recall of 0.98 at the baseline operating point. A physics-informed post-processing pipeline converts detections into two operationally distinct modes. The first is a high-sensitivity mode that applies morphological filtering and proximity-based merging for comprehensive emission screening, achieving precision of 0.71 and recall of 0.94. The second is a high-precision mode that additionally applies a distribution-based classifier for confident source attribution, achieving precision of 0.92 and recall of 0.70. Manual review of detections classified as false positives against our wavelet-based reference plume mask labels shows that some correspond to real methane enhancements... Our data and code used in this work are publicly available at: https://doi.org/10.7910/DVN/FR959H

cs.CV↗

arXiv2026

The Galaxy's Guide to the Tokenizer: A Benchmark for Scientific Foundation Models

Tokenization is central to adapting scientific data for transformer-based foundation models, yet its impact on learned representations remains poorly understood. We compare four tokenization strategies, Affine, AIM, JetFormer, and VQ-VAE, within a unified transformer framework for astronomical imaging. Using 640,000 galaxy images from the DESI Legacy Survey and a shared AstroPT backbone, we evaluate each method on reconstruction fidelity and prediction of physical properties. Our results reveal trade-offs across approaches. The flow-based JetFormer achieves higher reconstruction quality, while VQ-VAE yields strong probe performance for galaxy physical properties. Affine and AIM better preserve localized morphological information. We find that reconstruction and representation quality are decoupled, and no single method consistently performs best across the tasks considered here. By grounding our evaluation in independently measured physical quantities, we hope this study serves to highlight the potential of scientific data as a basis for constructing interpretable benchmarks for foundation models.

astro-ph.IM↗

arXiv2023

Hyperspectral shadow removal with Iterative Logistic Regression and latent Parametric Linear Combination of Gaussians

Shadow detection and removal is a challenging problem in the analysis of hyperspectral images. Yet, this step is crucial for analyzing data for remote sensing applications like methane detection. In this work, we develop a shadow detection and removal method only based on the spectrum of each pixel and the overall distribution of spectral values. We first introduce Iterative Logistic Regression (ILR) to learn a spectral basis in which shadows can be linearly classified. We then model the joint distribution of the mean radiance and the projection coefficients of the spectra onto the above basis as a parametric linear combination of Gaussians. We can then extract the maximum likelihood mixing parameter of the Gaussians to estimate the shadow coverage and to correct the shadowed spectra. Our correction scheme reduces correction artefacts at shadow borders. The shadow detection and removal method is applied to hyperspectral images from MethaneAIR, a precursor to the satellite MethaneSAT.

physics.data-an↗

arXiv2023

Domain Adaptation via Minimax Entropy for Real/Bogus Classification of Astronomical Alerts

Time domain astronomy is advancing towards the analysis of multiple massive datasets in real time, prompting the development of multi-stream machine learning models. In this work, we study Domain Adaptation (DA) for real/bogus classification of astronomical alerts using four different datasets: HiTS, DES, ATLAS, and ZTF. We study the domain shift between these datasets, and improve a naive deep learning classification model by using a fine tuning approach and semi-supervised deep DA via Minimax Entropy (MME). We compare the balanced accuracy of these models for different source-target scenarios. We find that both the fine tuning and MME models improve significantly the base model with as few as one labeled item per class coming from the target dataset, but that the MME does not compromise its performance on the source dataset.

astro-ph.IM↗

arXiv2023

Con$^{2}$DA: Simplifying Semi-supervised Domain Adaptation by Learning Consistent and Contrastive Feature Representations

In this work, we present Con$^{2}$DA, a simple framework that extends recent advances in semi-supervised learning to the semi-supervised domain adaptation (SSDA) problem. Our framework generates pairs of associated samples by performing stochastic data transformations to a given input. Associated data pairs are mapped to a feature representation space using a feature extractor. We use different loss functions to enforce consistency between the feature representations of associated data pairs of samples. We show that these learned representations are useful to deal with differences in data distributions in the domain adaptation problem. We performed experiments to study the main components of our model and we show that (i) learning of the consistent and contrastive feature representations is crucial to extract good discriminative features across different domains, and ii) our model benefits from the use of strong augmentation policies. With these findings, our method achieves state-of-the-art performances in three benchmark datasets for SSDA.

cs.LG↗

arXiv2023

Positional Encodings for Light Curve Transformers: Playing with Positions and Attention

We conducted empirical experiments to assess the transferability of a light curve transformer to datasets with different cadences and magnitude distributions using various positional encodings (PEs). We proposed a new approach to incorporate the temporal information directly to the output of the last attention layer. Our results indicated that using trainable PEs lead to significant improvements in the transformer performances and training times. Our proposed PE on attention can be trained faster than the traditional non-trainable PE transformer while achieving competitive results when transfered to other datasets.

astro-ph.IM↗

arXiv2023

Multi-Class Deep SVDD: Anomaly Detection Approach in Astronomy with Distinct Inlier Categories

With the increasing volume of astronomical data generated by modern survey telescopes, automated pipelines and machine learning techniques have become crucial for analyzing and extracting knowledge from these datasets. Anomaly detection, i.e. the task of identifying irregular or unexpected patterns in the data, is a complex challenge in astronomy. In this paper, we propose Multi-Class Deep Support Vector Data Description (MCDSVDD), an extension of the state-of-the-art anomaly detection algorithm One-Class Deep SVDD, specifically designed to handle different inlier categories with distinct data distributions. MCDSVDD uses a neural network to map the data into hyperspheres, where each hypersphere represents a specific inlier category. The distance of each sample from the centers of these hyperspheres determines the anomaly score. We evaluate the effectiveness of MCDSVDD by comparing its performance with several anomaly detection algorithms on a large dataset of astronomical light-curves obtained from the Zwicky Transient Facility. Our results demonstrate the efficacy of MCDSVDD in detecting anomalous sources while leveraging the presence of different inlier categories. The code and the data needed to reproduce our results are publicly available at https://github.com/mperezcarrasco/AnomalyALeRCE.

cs.LG↗

arXiv2021

Matching Embeddings for Domain Adaptation

In this work we address the problem of transferring knowledge obtained from a vast annotated source domain to a low labeled target domain. We propose Adversarial Variational Domain Adaptation (AVDA), a semi-supervised domain adaptation method based on deep variational embedded representations. We use approximate inference and domain adversarial methods to map samples from source and target domains into an aligned class-dependent embedding defined as a Gaussian Mixture Model. AVDA works as a classifier and considers a generative model that helps this classification. We used digits dataset for experimentation. Our results show that on a semi-supervised few-shot scenario our model outperforms previous methods in most of the adaptation tasks, even using a fewer number of labeled samples per class on target domain.

cs.LG↗

arXiv2018

Multiband galaxy morphologies for CLASH: a convolutional neural network transferred from CANDELS

We present visual-like morphologies over 16 photometric bands, from ultra-violet to near infrared, for 8,412 galaxies in the Cluster Lensing And Supernova survey with Hubble (CLASH) obtained by a convolutional neural network (CNN) model. Our model follows the CANDELS main morphological classification scheme, obtaining the probability for each galaxy at each CLASH band of being spheroid, disk, irregular, point source, or unclassifiable. Our catalog contains morphologies for each galaxy with Hmag < 24.5 in every filter where the galaxy is observed. We trained an initial CNN model using approximately 7,500 expert eyeball labels from The Cosmic Assembly Near-IR Deep Extragalactic Legacy Survey (CANDELS). We created eyeball labels for 100 randomly selected galaxies per each of the 16-filters set of CLASH (1,600 galaxy images in total), where each image was classified by at least five of us. We use these labels to fine-tune the network in order to accurately predict labels for the CLASH data and to evaluate the performance of our model. We achieve a root-mean-square error of 0.0991 on the test set. We show that our proposed fine-tuning technique reduces the number of labeled images needed for training, as compared to directly training over the CLASH data, and achieves a better performance. This approach is very useful to minimize eyeball labeling efforts when classifying unlabeled data from new surveys. This will become particularly useful for massive datasets such as the ones coming from near future surveys such as EUCLID or the LSST. Our catalog consists of prediction of probabilities for each galaxy by morphology in their different bands and is made publicly available at http://www.inf.udec.cl/~guille/data/Deep-CLASH.csv.

astro-ph.IM↗