High-dynamic-range (HDR) image reconstruction involves creating an HDR image from multiple low-dynamic-range images as input, providing a computational solution to enhance image quality. This task presents several challenges, such as frame misalignment, overexposure, and motion, which are addressed using deep learning algorithms. In this context, various architectures with different approaches exist, such as convolutional neural networks, diffusion networks, generative adversarial networks, and Transformer-based architectures, with the latter offering the best quality but at a high computational cost. This paper proposes an HDR reconstruction architecture using a Transformer-based approach to achieve results competitive with the state of the art while reducing computational cost. The number of self-attention blocks was reduced for feature refinement. To prevent quality degradation, a Convolutional Block Attention Module was added, enhancing image features by using the central frame as a reference. The proposed architecture was evaluated on two datasets, achieving the best results on Tel’s dataset in terms of quality metrics. The computational cost indicated that the architecture was significantly more efficient than other Transformer-based approaches for reconstruction. The results of this research suggest that low-complexity Transformer-based architectures have great potential, with applications extending beyond HDR reconstruction to other domains.

High Dynamic Range (HDR) image reconstruction has gained prominence in a wide range of fields; not only is it implemented in computer vision, but industries such as entertainment and medicine also benefit considerably from this technology due to its ability to capture and reproduce scenes with a greater variety of luminosities, extending conventional levels of perception. This article presents a review of the state of the art of HDR reconstruction methods based on deep learning, ranging from classical approaches that are still expressive and relevant to more recent proposals involving the advent of new architectures. The fundamental role of high-quality datasets and specific metrics in evaluating the performance of HDR algorithms is also discussed, as well as emphasizing the challenges inherent in capturing multiple exposures and dealing with artifacts. Finally, emerging trends and promising directions for overcoming current limitations and expanding the potential of HDR reconstruction in real-world scenarios are highlighted.

Printed circuit boards (PCBs) are key components in the electronics industry, and ensuring their integrity is essential for reliable manufacturing. Automated inspection systems based on computer vision, although efficient, face challenges. In this scenario, deep learning techniques have become effective solutions for detecting defects in more modern and complex PCBs. This article presents a comparative study between the YOLOv8n, YOLOv11n and RT-DETRv2 models for identifying defects in PCBs. The experiments were conducted using the PKU-Market-PCB dataset, which includes Missing Hole, Mouse Bite, Open Circuit, Short Circuit, Spur and Spurious Copper defects. To reduce the computational cost, modified versions of YOLOv8n and YOLOv11n with Depthwise convolution blocks (YOLOv8-DWConv and YOLOv11-DWConv). The analysis includes quantitative and qualitative comparisons. In addition, the robustness of the models is evaluated under challenging conditions with blur and illumination gradient noise. The results indicate that YOLOv11n achieves the best overall performance, while YOLOv11n-DWConv offers a competitive balance between precision and computational efficiency.

Multi-exposure high dynamic range (HDR) image reconstruction makes it possible to improve the quality of low dynamic range (LDR) images using algorithms based on deep learning. The main challenge in this task lies in the generation of ghosting artifacts in the resulting image, caused by the displacement of objects and by the low or overexposure in the input LDR images. To overcome this challenge, this paper proposes the FLASH-S architecture, which incorporates an innovative attention block using Vision Mamba state-space models (SSM) to enhance the extracted features from the input images. The features are serialized in four different directions to guarantee the same importance in each patch of the sequence in the SSM. In addition, the FLASH architecture was proposed, an improved version of FLASH-S using optical flow to align the input images. Experiments were carried out using the Kalantari, SynHDR and Tursun datasets. Based on seven quality metrics and three computational cost metrics, the results showed that the FLASH-S and FLASH proposals are superior to state-of-the-art algorithms. Additionally, FLASH demonstrated better reconstruction quality in qualitative evaluations, generalization capacity and tests in real environments. The model’s source code can be obtained from: https://github.com/pavic-lab-ufac/FLASH.