Search

The conventional pretraining-and-finetuning paradigm, while effective for common diseases with ample data, faces challenges in diagnosing data-scarce occupational diseases like pneumoconiosis. Recently, large language models (LLMs) have exhibits unprecedented ability when conducting multiple tasks in dialogue, bringing opportunities to diagnosis. A common strategy might involve using adapter layers for vision-language alignment and diagnosis in a dialogic manner. Yet, this approach often requires optimization of extensive learnable parameters in the text branch and the dialogue head, potentially diminishing the LLMs' efficacy, especially with limited training data. In our work, we innovate by eliminating the text branch and substituting the dialogue head with a classification head. This approach presents a more effective method for harnessing LLMs in diagnosis with fewer learnable parameters. Furthermore, to balance the retention of detailed image information with progression towards accurate diagnosis, we introduce the contextual multi-token engine. This engine is specialized in adaptively generating diagnostic tokens. Additionally, we propose the information emitter module, which unidirectionally emits information from image tokens to diagnosis tokens. Comprehensive experiments validate the superiority of our methods and the effectiveness of proposed modules. Our codes can be found at https://github.com/CodeMonsterPHD/PneumoLLM/tree/main., Comment: Medical Image Analysis

Backprojection networks have achieved promising super-resolution performance for nature images but not well be explored in the remote sensing image super-resolution (RSISR) field due to the high computation costs. In this paper, we propose a Multi-granularity Backprojection Transformer termed MBT for RSISR. MBT incorporates the backprojection learning strategy into a Transformer framework. It consists of Scale-aware Backprojection-based Transformer Layers (SPTLs) for scale-aware low-resolution feature learning and Context-aware Backprojection-based Transformer Blocks (CPTBs) for hierarchical feature learning. A backprojection-based reconstruction module (PRM) is also introduced to enhance the hierarchical features for image reconstruction. MBT stands out by efficiently learning low-resolution features without excessive modules for high-resolution processing, resulting in lower computational resources. Experiment results on UCMerced and AID datasets demonstrate that MBT obtains state-of-the-art results compared to other leading methods.

We introduce a framework to measure the asphericity of Sun-like stars using $a_1$, $a_2$ and $a_4$ coefficients, and constrain their latitudes of magnetic activity. Systematic errors on the inferred coefficients are evaluated in function of key physical and seismic parameters (inclination of rotation axis, average rotation, height-to-noise ratio of peaks in power spectrum). The measured a-coefficients account for rotational oblateness and the effect of surface magnetic activity. We use a simple model that assumes a single latitudinal band of activity. Using solar SOHO/VIRGO/SPM data, we demonstrate the capability of the method to detect the mean active latitude and its intensity changes between 1999-2002 (maximum of activity) and 2006-2009 (minimum of activity). We further apply the method to study the solar-analogue stars 16 Cyg A and B using Kepler observations. An equatorial band of activity, exhibiting intensity that could be comparable to that of the Sun, is detected in 16 Cyg A. However, 16 Cyg B exhibits a bi-modality in $a_4$ that is challenging to explain. We suggest that this could be a manifestation of the transition between a quiet and an active phase of activity. Validating or invalidating this hypothesis may require new observations., Comment: Accepted to A&A

Remote sensing image super-resolution (RSISR) plays a vital role in enhancing spatial detials and improving the quality of satellite imagery. Recently, Transformer-based models have shown competitive performance in RSISR. To mitigate the quadratic computational complexity resulting from global self-attention, various methods constrain attention to a local window, enhancing its efficiency. Consequently, the receptive fields in a single attention layer are inadequate, leading to insufficient context modeling. Furthermore, while most transform-based approaches reuse shallow features through skip connections, relying solely on these connections treats shallow and deep features equally, impeding the model's ability to characterize them. To address these issues, we propose a novel transformer architecture called Cross-Spatial Pixel Integration and Cross-Stage Feature Fusion Based Transformer Network (SPIFFNet) for RSISR. Our proposed model effectively enhances global cognition and understanding of the entire image, facilitating efficient integration of features cross-stages. The model incorporates cross-spatial pixel integration attention (CSPIA) to introduce contextual information into a local window, while cross-stage feature fusion attention (CSFFA) adaptively fuses features from the previous stage to improve feature expression in line with the requirements of the current stage. We conducted comprehensive experiments on multiple benchmark datasets, demonstrating the superior performance of our proposed SPIFFNet in terms of both quantitative metrics and visual quality when compared to state-of-the-art methods.

We perform photometric (the Lomb-Scargle periodogram, autocorrelation, and wavelet) and asteroseismic analyses of 92 Kepler solar-like main-sequence stars to understand the reliability of the measured stellar rotation periods. We focus on the 70 stars without reported stellar companions, and classify them into four groups according to the quarter-to-quarter variance of the Lomb-Scargle period and the precision of the asteroseismic period. We present detailed individual comparison among photometric and asteroseismic constraints for these stars. We find that most of our targets exhibit significant quarter-to-quarter variances in the photometric periods, suggesting that the photometrically estimated period should be regarded as a simplified characterization of the true stellar rotation period, especially under the presence of the latitudinal differential rotation. On the other hand, there are a fraction of stars with a relatively small quarter-to-quarter variance in the photometric periods, most of which have consistent values for asteroseismically and photometrically estimated rotation periods. We also identify over ten stars whose photometric and asteroseismic periods significantly disagree, which would be potentially interesting targets for further individual investigations.

This work theoretically studies stochastic neural networks, a main type of neural network in use. We prove that as the width of an optimized stochastic neural network tends to infinity, its predictive variance on the training set decreases to zero. Our theory justifies the common intuition that adding stochasticity to the model can help regularize the model by introducing an averaging effect. Two common examples that our theory can be relevant to are neural networks with dropout and Bayesian latent variable models in a special limit. Our result thus helps better understand how stochasticity affects the learning of neural networks and potentially design better architectures for practical problems.

Most finger vein feature extraction algorithms achieve satisfactory performance due to their texture representation abilities, despite simultaneously ignoring the intensity distribution that is formed by the finger tissue, and in some cases, processing it as background noise. In this paper, we exploit this kind of noise as a novel soft biometric trait for achieving better finger vein recognition performance. First, a detailed analysis of the finger vein imaging principle and the characteristics of the image are presented to show that the intensity distribution that is formed by the finger tissue in the background can be extracted as a soft biometric trait for recognition. Then, two finger vein background layer extraction algorithms and three soft biometric trait extraction algorithms are proposed for intensity distribution feature extraction. Finally, a hybrid matching strategy is proposed to solve the issue of dimension difference between the primary and soft biometric traits on the score level. A series of rigorous contrast experiments on three open-access databases demonstrates that our proposed method is feasible and effective for finger vein recognition., Comment: 11 pages