Single-image dehazing is a pivotal challenge in computer vision that seeks to remove haze from images and restore clean background details. Recognizing the limitations of traditional physical model-based methods and the inefficiencies of current attention-based solutions, we propose a new dehazing network combining an innovative Haze-Aware Attention Module (HAAM) with a Multiscale Frequency Enhancement Module (MFEM). The HAAM is inspired by the atmospheric scattering model, thus skillfully integrating physical principles into high-dimensional features for targeted dehazing. It picks up on latent features during the image restoration process, which gives a significant boost to the metrics, while the MFEM efficiently enhances high-frequency details, thus sidestepping wavelet or Fourier transform complexities. It employs multiscale fields to extract and emphasize key frequency components with minimal parameter overhead. Integrated into a simple U-Net framework, our Haze-Aware Attention Network (HAA-Net) for single-image dehazing significantly outperforms existing attention-based and transformer models in efficiency and effectiveness. Tested across various public datasets, the HAA-Net sets new performance benchmarks. Our work not only advances the field of image dehazing but also offers insights into the design of attention mechanisms for broader applications in computer vision.

The objective of single image dehazing is to restore hazy images and produce clear, high-quality visuals. Traditional convolutional models struggle with long-range dependencies due to their limited receptive field size. While Transformers excel at capturing such dependencies, their quadratic computational complexity in relation to feature map resolution makes them less suitable for pixel-to-pixel dense prediction tasks. Moreover, fixed kernels or tokens in most models do not adapt well to varying blur sizes, resulting in suboptimal dehazing performance. In this study, we introduce a novel dehazing network based on Parallel Stripe Cross Attention (PCSA) with a multi-scale strategy. PCSA efficiently integrates long-range dependencies by simultaneously capturing horizontal and vertical relationships, allowing each pixel to capture contextual cues from an expanded spatial domain. To handle different sizes and shapes of blurs flexibly, We employs a channel-wise design with varying convolutional kernel sizes and strip lengths in each PCSA to capture context information at different scales.Additionally, we incorporate a softmax-based adaptive weighting mechanism within PCSA to prioritize and leverage more critical features. We perform extensive testing on synthetic and real-world datasets and our PCSA-Net has established a new performance benchmark. This work not only advances the field of image dehazing but also serves as a reference for developing more diverse and efficient attention mechanisms. CCS Concepts: · Theory of computation → Theory and algorithms for application domains → Machine learning theory → Models of learning keywords: Parallel Cross Strip Attention, Single Image Dehazing, Horizontal Strip Attention, Vertical Strip Attention

In this paper, we introduced NanoTrack, a novel multi-object tracking (MOT) method that leverages light-weight object detectors to enhance tracking performance in real-time applications where computational resources are scarce. While light-weight detectors are efficient, they often produce an imbalance in detection quality, generating a significant number of low-scoring detections that pose challenges for tracking algorithms. Our approach innovatively utilizes these low-scoring detections for track initialization and maintenance, addressing the shortcomings observed in existing tracking by two-stage tracking methods like ByteTrack, which struggle with the abundance of low-scoring detections. By integrating two new light-weight modules, Refind High Detection (RHD) and Duplicate Track Checking (DTC), NanoTrack effectively incorporates low-scoring detections into the tracking process. Additionally, we enhance the pseudo-depth estimation technique for improved handling in dense target environments, mitigating issues like ID Switching. Our comprehensive experiments demonstrate that NanoTrack surpasses state-of-the-art two-stage TBD methods, including ByteTrack and SparseTrack, on benchmark datasets such as MOT16, MOT17, and MOT20, thereby establishing a new standard for MOT performance using light-weight detectors. The code is open source in https://github.com/VjiaLi/NanoTrack