Image dehazing is a crucial low-level restoration task that aims to recover a clear image from a hazy observation. Recent learning-based approaches have demonstrated impressive performance in this area. However, there are still two drawbacks: (1) Existing dehazing architectures do not sufficiently consider non-uniform haze distribution, indicating that the haze location information is underexplored. (2) Naïve contrastive regularization fails to provide enough constraint force in solution space , i.e., negative-oriented supervision information cannot be fully utilized during the training stage. Consequently, we establish a Multi-stage Dehazing Network (MSD-Net) to achieve single image haze removal. For one thing, we build a haze perception unit (HPU) based on a self-calibration attentive paradigm. This unit can effectively encode haze location information as prior guidance and further enhance its feature representation capabilities. For another, we tailor a global and local progressive contrastive regularization (GLPCR) to explore negative-oriented supervision information. Specifically, the negative samples are derived not only from the original hazy images but are also progressively updated through the pseudo-restoration results of the multi-stage architecture. To tackle the learning ambiguity arising from diverse negative samples, we employ a curriculum learning strategy during the training phase. Moreover, our GLPCR operates in both global and local manners, encouraging the network to retain rich information from both image-wise and patch-wise perspectives. Extensive experiments demonstrate that our MSD-Net can achieve remarkable dehazing performance compared with other state-of-the-art methods on several commonly used hazy dataset benchmarks.

Point spread function (PSF) characterizes the intensity distribution characteristics of each object point and is widely used in areas such as defocus estimation, non-blind image deblurring, and computational imaging. Estimating spatially varying PSF from a single image is a typical inverse problem, which is constrained by multiple factors such as sensor noise and semantic information interference. In this paper, we propose a polynomial fitting-based method to model spatially varying PSF. With this method, we generate a large-scale, high-quality dataset with pixel-level annotations that can be used for training deep learning networks. To solve the task of estimating defocus maps from a single image, we design a novel high-resolution coefficient regression network to achieve accurate defocus estimation and concurrent estimation of multiple aberrations, respectively. To the best of our knowledge, this work presents the inaugural attempt at spatially varying PSF estimation based on polynomial coefficient regression. Extensive experimental results show that our methodology attains state-of-the-art performance across numerous evaluation metrics, fully verifying its effectiveness and superiority. The dataset and code is available on GitHub: https://github.com/67689E4F/PSFNet.git.

Visible and infrared image registration is a challenging problem in computer vision due to the significant differences in appearance and physical properties between the two modalities. A single feature is not enough to remove nonlinear differences, and the matching method faces a trade-off between the high-resolution feature map and the transformer model. In this paper, we propose a novel method called Adaptive-Neighborhood Contextual Multi-features (ANCM-Net) for visible/infrared image registration. Our method addresses the limitations of existing approaches by incorporating depth features and cross-modal similar contour features to form contextual feature representations. Additionally, we propose a region-spanning adaptive cross-attention module to handle low spatial resolution and redundancy in attention computation. This module enables attentional encoding of limited information in the attention location and cross-modal adaptive region through attention region adjustment. In the matching task, we compute an adaptive attention region for each pixel point in the cross-modal image and encode and match the depth features and edge features together. As a result, ANCM-Net not only preserves the long-range dependency of the image feature structure but also achieves fine-grained attention between highly correlated pixels. By extracting cross-modal consistent contextual features to compensate for modality-specific information, our approach improves the cross-modal matching performance. Extensive experiments on real-world captured thermal infrared and visible datasets demonstrate that ANCM-Net outperforms existing image matching methods.

The performance of visual classification models across Internet of Things devices is usually limited by the changes in local environments, resulted from the diverse appearances of the target objects and differences in light conditions and background scenes. To alleviate these problems, existing studies usually introduce the multimodal information to guide the learning process of the visual classification models, making the models extract the visual features from the discriminative image regions. Especially, cross-modal alignment between visual and textual features has been considered as an effective way for this task by learning a domain-consistent latent feature space for the visual and semantic features. However, this approach may suffer from the heterogeneity between multiple modalities, such as the multimodal features and the differences in the learned feature values. To alleviate this problem, this article first presents a comparative analysis of the functionality of various alignment strategies and their impacts on improving visual classification. Subsequently, a cross-modal inference and fusion framework (termed as CRIF) is proposed to align the heterogeneous features in both the feature distributions and values. More importantly, CRIF includes a cross-modal information enrichment module to improve the final classification and learn the mappings from the visual to the semantic space. We conduct experiments on four benchmarking data sets, i.e., the Vireo-Food172, NUS-WIDE, MSR-VTT, and ActivityNet Captions data sets. We report state-of-the-art results for basic classification tasks on the four data sets and conduct subsequent experiments on feature alignment and fusion. The experimental results verify that CRIF can effectively improve the learning ability of the visual classification models, and it is a model-agnostic framework that consistently improves the performance of state-of-the-art visual classification models.

In 2009, a 12-week-long project-based course Optoelectronic Instrument Experiments (OIE) was launched at the School of Opto-Electronics, Beijing Institute of Technology. During the classes, the students are assigned to teams and each team chooses one project to implement in twelve weeks. Through the mini simulative "Cycle of Professional Practice", students will learn how to integrate the knowledge and techniques they have learned previously, and use different kinds of components and devices to construct an optoelectronic instrument. The main difference from the other project-based courses is that the OIE course is compulsory, which means every student must take this course. Here comes the problem. Young students usually are rebellious somehow. They would be enthusing about what they choose to do, while would be reluctant to even the same thing if they were arranged to do it. Moreover, students from China are good at theoretical knowledge but lack of initiative. Most of them prefer working alone to being a team member. These are stereotypes but also partly truth. Inactivity with repellent mood and initial resistance to team-based approaches became the biggest barrier in obligatory project-based courses. Aiming to solve those issues, several approaches were tried out during the OIE course design and the progresses, including the design of based projects, aptitude digging process and teamwork pedagogy, which seemed promising and inspiring in both stimulating students' enthusiasm and encouraging their team spirits. In this paper, we will give the detail of our tryout and analyses. The OIE course has been carried out for five academic years by now. Surveys conducted among the students who have taken the course and analyses are also included.

Students from China are always intelligent but lack of creativity. These are somewhat stereotypes. This is partially because of the reserved or implicit culture. In an objective point of view, it is also because of the limited education resources. In the single assessment criterion education circumstance, students chase for the high marks even without knowing their interests or aptitudes. In a 12-week open experimental course, Optoelectronic Instrument Experiments (OIE), we try to encourage the students to dig their aptitudes and bring them into full play to earn more credits for the course. Self-assessment and mutual-evaluation for technical proficiency, communication skills, collaboration and leadership are carried out for the final evaluation. We also communicate with the students the speciality and skill a qualified engineer needs. We hope to help them prepare themselves for engineering-related jobs in the further.

Project-based course is based on teamwork and most of work is done and presented as a team. Team grouping rule is one of the most important issues. In project-based experimental course Optoelectronic Instrument Experiments (OIE), several different rules were attempted, each of which produced complaints by some students. After several trials of different grouping rule, we realized that there is no perfect rule which can satisfy everyone. Instead of changing the rule, trying to find a way to persuade the students to accept and support their group willingly might be a better solution. In this semester, the project teams are entirely determined by lot and several teamwork inspirational approaches are introduced to inspire team spirit in the course. Our purpose is to find a way to make student learn the interpersonal skill of working in team. Let's do it, not just deal with it inactively.