PlateUNLP is a semi-automatic software tool designed to assist in the processing of digitized spectral plate images. It automates key stages such as spectra detection, spectral extraction, wavelength calibration, and metadata generation. These tasks, traditionally performed manually, are significantly accelerated while still allowing user intervention when necessary. This work focuses on the integration of automation for the detection of spectra in images using deep learning techniques, as well as the extraction of one-dimensional spectra based on the image segments corresponding to each spectrum, and their subsequent wavelength calibration. The tool is being developed in collaboration with expert astronomers, demonstrating high reliability and practical utility in the recovery of historical astronomical data.

In this study, we examine the potential of using high-resolution satellite imagery and machine learning techniques to create income maps with a high level of geographic detail. We trained a convolutional neural network with satellite images from the Metropolitan Area of Buenos Aires (Argentina) and 2010 census data to estimate per capita income at a 50x50 meter resolution for 2013, 2018, and 2022. This outperformed the resolution and frequency of available census information. Based on the MobileNetV3 architecture, the model achieved high accuracy in predicting household incomes (R2= 0.77), surpassing the spatial resolution and model performance of other methods used in the existing literature. This approach presents new opportunities for the generation of highly disaggregated data, enabling the assessment of public policies at a local scale, providing tools for better targeting of social programs, and reducing the information gap in areas where data is not collected.

The performance of diagnostic Computer-Aided Design (CAD) systems for retinal diseases depends on the quality of the retinal images being screened. Thus, many studies have been developed to evaluate and assess the quality of such retinal images. However, most of them did not investigate the relationship between the accuracy of the developed models and the quality of the visualization of interpretability methods for distinguishing between gradable and non-gradable retinal images. Consequently, this paper presents a novel framework called “FGR-Net” to automatically asses and interpret underlying fundus image quality by merging an autoencoder network with a classifier network. The FGR-Net model also provides an interpretable quality assessment through visualizations. In particular, FGR-Net uses a deep autoencoder to reconstruct the input image in order to extract the visual characteristics of the input fundus images based on self-supervised learning. The extracted features by the autoencoder are then fed into a deep classifier network to distinguish between gradable and ungradable fundus images. FGR-Net is evaluated with different interpretability methods, which indicates that the autoencoder is a key factor in forcing the classifier to focus on the relevant structures of the fundus images, such as the fovea, optic disc, and prominent blood vessels. Additionally, the interpretability methods can provide visual feedback for ophthalmologists to understand how our model evaluates the quality of fundus images. The experimental results showed the superiority of FGR-Net over the state-of-the-art quality assessment methods, with an accuracy of >89% and an F1-score of >87%. The code is publicly available at https://github.com/saifalkh/FGR-Net.

Abstract Widespread eye conditions such as cataracts, diabetic retinopathy, and glaucoma impact people worldwide. Ophthalmology uses fundus photography for diagnosing these retinal disorders, but fundus images are prone to image quality challenges. Accurate diagnosis hinges on high-quality fundus images. Therefore, there is a need for image quality assessment methods to evaluate fundus images before diagnosis. Consequently, this paper introduces a deep learning model tailored for fundus images that supports large images. Our division method centres on preserving the original image’s high-resolution features while maintaining low computing and high accuracy. The proposed approach encompasses two fundamental components: an autoencoder model for input image reconstruction and image classification to classify the image quality based on the latent features extracted by the autoencoder, all performed at the original image size, without alteration, before reassembly for decoding networks. Through post hoc interpretability methods, we verified that our model focuses on key elements of fundus image quality. Additionally, an intrinsic interpretability module has been designed into the network that allows decomposing class scores into underlying concepts quality such as brightness or presence of anatomical structures. Experimental results in our model with EyeQ, a fundus image dataset with three categories (Good, Usable, and Rejected) demonstrate that our approach produces competitive outcomes compared to other deep learning-based methods with an overall accuracy of 0.9066, a precision of 0.8843, a recall of 0.8905, and an impressive F1-score of 0.8868. The code is publicly available at https://github.com/saifalkhaldiurv/VISTA_-Image-Quality-Assessment .

As the presence of artificial intelligence has increased in a variety of different areas, the use of machine learning and deep learning techniques for creative purposes has also risen significantly in recent years. Works of this kind within the area of natural language processing (NLP) are typically neural models used for fiction or lyrics generation. Those works are in most cases in English and adapting them to other languages is not feasible. In this work, we develop a Spanish text generator system for the rap sub-genre known as freestyle. Freestyle songs present unique challenges for text generation given that performers compete with one another in a lyric improvisation contest. Given the low availability of freestyle text, especially in Spanish, we collected two separate datasets, one with freestyle lyrics and the other, larger, with rap lyrics, which are more readily available. The rap dataset can be used for pretraining, and the freestyle dataset for finetuning on the generation task. Furthermore, we design a neural network-based generation model that takes into account both the structure of freestyle and the low data availability. The model was able to generate realistic freestyle verses in Spanish.