Large and diverse image datasets have facilitated the recent advances in deep-learning-based computer vision applications. Whereas datasets with images depicting normal-weather scenes are plentiful, datasets with images depicting inclement weather conditions, such as haze, remain scarce due to collection difficulties. In response to this problem, we present a novel domain flow adaptation network (DFA-Net) that can control the haze density and facilitate the generation of realistic and diverse hazy images. DFA-Net employs a density variable to direct the network to learn and yield the desired images and is composed of four modules: a semantic extraction (SE) module, a haze extraction (HE) module, an image production (IP) module, and an image assessment (IS) module. The SE and HE modules are used to capture the semantic structure and style representation of clear and hazy images, respectively, and provide them to the IP module for refining the output images. The IP module is adopted to yield hazy images in a coarse-to-fine fashion, while the IS module is responsible for examining the realism of the synthesized results. Experiments on multiple benchmark datasets confirm the effectiveness of the proposed DFA-Net, which outperforms competing approaches by achieving improvements of up to 147% in quality, 237% in fidelity, and 354% in the diversity of generated images.

Large-scale datasets centered on images have driven advancements in deep learning-based computer vision applications. While there is an abundance of datasets containing images depicting favorable weather scenes, datasets featuring images of adverse weather conditions, especially the presence of haze, are scarce due to challenges in their collection. In response, we leverage the advantages of deep learning techniques to introduce a novel approach for facilitating the rendering of realistic and diverse hazy images, named RenHaze. To be specific, RenHaze adopts a denseness parameter \(\omega\) to control the haze level of output images and consists of five subnets, including a content exploitation (CE) subnet, a depth exploitation (DE) subnet, a haze exploitation (HE) subnet, an image generation (IG) subnet, and an image discernment (ID) subnet. The CE, DE, and HE subnets are responsible for extracting features from the source clear image, depth image, and reference hazy image, respectively, and then providing them for the IG subnet. The IG subnet is used to perform image translation in a coarse-to-fine manner, while the ID subnet is employed to discern the realism of the rendered image and provide feedback to the IG subnet for generating the desired output. Extensive experiments demonstrate the superiority of the proposed model over competing IG methods in terms of the realism and diversity of synthesized hazy images, as well as its effectiveness in boosting the performance of computer vision tasks such as object detection and semantic segmentation in real-world hazy environments.

In recent years, object detection has significantly advanced by using deep learning, especially convolutional neural networks. Most of the existing methods have focused on detecting objects under favorable weather conditions and achieved impressive results. However, object detection in the presence of rain remains a crucial challenge owing to the visibility limitation. In this article, we introduce an amalgamating knowledge network (AK-Net) to deal with the problem of detecting objects hampered by rain. The proposed AK-Net obtains performance improvement by associating object detection with visibility enhancement, and it is composed of five subnetworks: rain streak removal (RSR) subnetwork, raindrop removal (RDR) subnetwork, foggy rain removal (FRR) subnetwork, feature transmission (FT) subnetwork, and object detection (OD) subnetwork. Our approach is flexible; it can adopt different object detection models to construct the OD subnetwork for the final inference of objects. The RSR, RDR, and FRR subnetworks are responsible for producing clean features from rain streak, raindrop, and foggy rain images, respectively, and offer them to the OD subnetwork through the FT subnetwork for efficient object prediction. Experimental results indicate that the mean average precision (mAP) achieved by our proposed AK-Net was up to 19.58% and 26.91% higher than those produced using competitive methods on published iRain and RID datasets, respectively, while preserving the fast-running time of the baseline detector.

In recent years, deep convolutional neural networks (CNNs) have been widely applied and have gained considerable success in object detection (OD). However, most of the CNN-based object detectors have been developed to operate under favorable weather conditions, limiting their ability to accurately detect objects in rainy nighttime (RNT) scenes, thereby resulting in low performance. In this work, we introduce a multilevel knowledge transmission network (MKT-Net) to overcome the challenges of detecting objects with the interference of rain and night. Our proposed model accomplishes this objective by collaborating OD with rain removal (RR) and low-illumination enhancement (LE) tasks. Specifically, the MKT-Net is composed of three main subnetworks that share some shallow layers with each other: an OD subnetwork for performing object classification and localization, an RR subnetwork, and an LE subnetwork for generating clear features. To aggregate and transmit multiscale features generated by the RR and LE subnetworks to the OD subnetwork for boosting detection accuracy, we introduce two feature transmission modules with identical architectures. Extensive evaluation on various datasets has demonstrated the effectiveness of our proposed model, which outperformed competing methods by up to 25.43% and 15.26% in mean average precision on a collected RNT dataset and the published rain in driving dataset, respectively, while maintaining high detection speed.

Numerous lightweight detection models have been presented in recent years, yet these detectors are inclined to develop for operating under normal weather conditions without adequate studies for rainy conditions. This is one of the causes leads drastically performance degradation of object detectors due to the decrease in visibility. To address above insufficiency, we propose a new and effective approach, named 3FL-Net, to elevate the performance of lightweight object detectors in the presence of rain. Our approach fulfills the goal by closely incorporating four subnetworks, namely feature enhancement subnetwork, feature extraction subnetwork, feature adaptation subnetwork, and lightweight detection subnetwork. The lightweight detection subnetwork achieved the accuracy improvement by learning diverse features from the feature enhancement subnetwork and feature extraction subnetwork via the feature adaptation subnetwork. To further drive the development in object detection induced by rain, we introduce a large-scale driving dataset, called iRain. The full iRain consists of 17,950 real-world rain images, which covers most of the driving scenarios and 85,081 instances explaining five prevalent object classes. Experiment results on divergent rain datasets expose that our 3FL-Net considerably improves the performance of lightweight detectors and surpasses that of the combination models between rain removal and object detection methods.

In recent years, object detection approaches using deep convolutional neural networks (CNNs) have derived major advances in normal images. However, such success is hardly achieved with rainy images due to lack of visibility. Aiming to bridge this gap, in this article, we present a novel selective features absorption network (SFA-Net) to improve the performance of object detection not only in rainy weather conditions but also in favorable weather conditions. SFA-Net accomplishes this objective by utilizing three subnetworks, where the feature selection subnetwork is concatenated with the object detection subnetwork through the feature absorption subnetwork to form a unified model. To promote further advancement in object detection impaired by rain, we propose a large-scale rainy image dataset, named srRain, which contains both synthetic rainy images and real-world rainy images for training and testing purposes. srRain is comprised of 25 900 rainy images depicting diverse driving scenarios in the presence of rain with a total of 181 164 instances interpreting five common object categories. Experimental results display that our SFA-Net reaches the highest mean average precision (mAP) of 77.53% on a normal image set, 62.52% on a synthetic rainy image set, 37.34% on a collected natural rainy image set, and 32.86% on a published real rainy image set, surpassing current state-of-the-art object detectors and the combination of image deraining and object detection models while retaining a high speed.

In the past half of the decade, object detection approaches based on the convolutional neural network have been widely studied and successfully applied in many computer vision applications. However, detecting objects in inclement weather conditions remains a major challenge because of poor visibility. In this article, we address the object detection problem in the presence of fog by introducing a novel dual-subnet network (DSNet) that can be trained end-to-end and jointly learn three tasks: visibility enhancement, object classification, and object localization. DSNet attains complete performance improvement by including two subnetworks: detection subnet and restoration subnet. We employ RetinaNet as a backbone network (also called detection subnet), which is responsible for learning to classify and locate objects. The restoration subnet is designed by sharing feature extraction layers with the detection subnet and adopting a feature recovery (FR) module for visibility enhancement. Experimental results show that our DSNet achieved 50.84 percent mean average precision (mAP) on a synthetic foggy dataset that we composed and 41.91 percent mAP on a public natural foggy dataset (Foggy Driving dataset), outperforming many state-of-the-art object detectors and combination models between dehazing and detection methods while maintaining a high speed.