LMHaze: Intensity-aware Image Dehazing with a Large-scale Multi-intensity Real Haze Dataset
This work addresses the challenge of real-world image dehazing for computer vision applications by providing a more diverse dataset and a novel model, though it is incremental in improving existing methods.
The authors tackled the problem of limited generalization in image dehazing due to dataset biases by introducing LMHaze, a large-scale real-world dataset with over 5K paired images spanning multiple haze intensities, and proposed a mixture-of-experts model based on Mamba that dynamically adjusts parameters for intensity-aware dehazing, achieving better results than state-of-the-art methods.
Image dehazing has drawn a significant attention in recent years. Learning-based methods usually require paired hazy and corresponding ground truth (haze-free) images for training. However, it is difficult to collect real-world image pairs, which prevents developments of existing methods. Although several works partially alleviate this issue by using synthetic datasets or small-scale real datasets. The haze intensity distribution bias and scene homogeneity in existing datasets limit the generalization ability of these methods, particularly when encountering images with previously unseen haze intensities. In this work, we present LMHaze, a large-scale, high-quality real-world dataset. LMHaze comprises paired hazy and haze-free images captured in diverse indoor and outdoor environments, spanning multiple scenarios and haze intensities. It contains over 5K high-resolution image pairs, surpassing the size of the biggest existing real-world dehazing dataset by over 25 times. Meanwhile, to better handle images with different haze intensities, we propose a mixture-of-experts model based on Mamba (MoE-Mamba) for dehazing, which dynamically adjusts the model parameters according to the haze intensity. Moreover, with our proposed dataset, we conduct a new large multimodal model (LMM)-based benchmark study to simulate human perception for evaluating dehazed images. Experiments demonstrate that LMHaze dataset improves the dehazing performance in real scenarios and our dehazing method provides better results compared to state-of-the-art methods.