Image dehazing on a hazy image aims to remove the haze and make the image scene clear, which attracts more and more research interests in recent years. Most existing image dehazing methods use a classic atmospheric scattering model and natural image priors to remove the image haze. In this paper, we propose an end-to-end image dehazing model termed as DRHNet (Deep Residual Haze Network), which restores the haze-free image by subtracting the learned negative residual map from the hazy image. Specifically, DRHNet proposes a context-aware feature extraction module to aggregate the contextual information effectively. Furthermore, it proposes a novel nonlinear activation function termed as RPReLU (Reverse Parametric Rectified Linear Unit) to improve its representation ability and to accelerate its convergence. Extensive experiments demonstrate that DRHNet outperforms state-of-the-art methods both quantitatively and qualitatively. In addition, experiments on image deraining task show that DRHNet can also serve for image deraining.

Enhancing the accuracy of dense classification with limited labeled data and abundant unlabeled data, known as semi-supervised semantic segmentation, is an essential task in vision comprehension. Due to the lack of annotation in unlabeled data, additional pseudo-supervised signals, typically pseudo-labeling, are required to improve the performance. Although effective, these methods fail to consider the internal representation of neural networks and the inherent class-imbalance in dense samples. In this work, we propose an information transfer theory, which establishes a theoretical relationship between shallow and deep representations. We further apply this theory at both the semantic and pixel levels, referred to as IIT-SP, to align different types of information. The proposed IIT-SP optimizes shallow representations to match the target representation required for segmentation. This limits the upper bound of deep representations to enhance segmentation performance. We also propose a momentum-based Cluster-State bar that updates class status online, along with a HardClassMix augmentation and a loss weighting technique to address class imbalance issues based on it. The effectiveness of the proposed method is demonstrated through comparative experiments on PASCAL VOC and Cityscapes benchmarks, where the proposed IIT-SP achieves state-of-the-art performance, reaching mIoU of 68.34% with only 2% labeled data on PASCAL VOC and mIoU of 64.20% with only 12.5% labeled data on Cityscapes.

Medical image segmentation has made significant progress when a large amount of labeled data are available. However, annotating medical image segmentation datasets is expensive due to the requirement of professional skills. Additionally, classes are often unevenly distributed in medical images, which severely affects the classification performance on minority classes. To address these problems, this paper proposes Co-Distribution Alignment (Co-DA) for semi-supervised medical image segmentation. Specifically, Co-DA aligns marginal predictions on unlabeled data to marginal predictions on labeled data in a class-wise manner with two differently initialized models before using the pseudo-labels generated by one model to supervise the other. Besides, we design an over-expectation cross-entropy loss for filtering the unlabeled pixels to reduce noise in their pseudo-labels. Quantitative and qualitative experiments on three public datasets demonstrate that the proposed approach outperforms existing state-of-the-art semi-supervised medical image segmentation methods on both the 2D CaDIS dataset and the 3D LGE-MRI and ACDC datasets, achieving an mIoU of 0.8515 with only 24% labeled data on CaDIS, and a Dice score of 0.8824 and 0.8773 with only 20% data on LGE-MRI and ACDC, respectively.

The number, relative ratio, and appearance of peripheral blood leukocytes can assist doctors to diagnose diseases such as lymphoma and leukemia. Therefore, segmentation of peripheral blood leukocytes in blood smear images plays a crucial role in the diagnosis of certain diseases. We propose a semisupervised leukocyte segmentation method under the framework of adversarial learning, in which a discriminator is trained to differentiate the segmentation maps coming either from the ground truth or from the segmentation network. Specifically, we first construct a lightweight leukocyte segmentation network by modifying DeepLabV3+ with simplified MobileNetV2. In addition, a reconstruction decoder is developed as a constraint on feature extraction to recover the original leukocyte image from its segmentation result. Then, we design the discriminator in a fully convolutional manner, which as well enables semi-supervised learning by inferring reliable supervisory information for unlabeled images. Quantitative and qualitative experimental results on four image datasets demonstrate the superiority of the proposed method over several state-of-the-art methods, achieving the highest average $F1$ score of 0.9717 and mIoU of 0.9493 with only 20% labeled training samples.