Single image-based dehazing has achieved remarkable progress with the development of deep learning technologies. End-to-end neural networks have been proposed to learn a direct hazy-to-clear image translation to recover the clear structures and edges cues from the hazy inputs. However, the frequency domain information is explored insufficiently and lots of intermediate structure and texture related cues of current dehazing networks are ignored, which limits the performances of current approaches. To handle these limitations mentioned above, a wavelet spatial attention based multi-stream feedback network (WSAMF-Net) is proposed for effective single image dehazing. Specifically, the proposed wavelet spatial attention utilizes both frequency-domain and spatial-domain information to enhance the extracted features for better structures and edges. Meanwhile, an enhanced multi-stream based cross feature fusion strategy, including vertical and horizontal attentions, is proposed to reweight and fuse the intermediate features of each stream to acquire more meaningful aggregated features, while the weight sharing strategy is used to achieve a good trade-off between performance and parameters. Besides, feedback mechanism is also designed to provide strong reconstruction ability. Furthermore, we propose a critical real-world industrial dataset (IDS) with images captured in real-world industrial quarry scenarios for research uses. Extensive experiments on various benchmarking datasets, including both synthetic and real-world datasets, demonstrate the superiority of our WSAMF-Net over state-of-the-art single image dehazing methods. The IDS dataset will be available at https://github.com/XBSong/IDS-Dataset.

Accurate segmentation and analysis for each animal in surveillance video images will help poultry farmers to monitor and promote animal welfare. However, it is challenging to accurately segment each animal due to the similar appearance, different scales, rapid growth and adhesive areas of group animals. Meanwhile, lacking of useful training data also limits the effectiveness of animal segmentation algorithms. To address these problems, we first construct a chicken image segmentation dataset to study the behavior of chickens for intelligent monitoring and analysis. Then, we propose an effective end-to-end framework for chicken image segmentation, which can also be used for other animal image segmentation. An end-to-end multi-scale based encoder-decoder network is first utilized to extract multi-scale features. Then, an attention-based module is employed to extract and intensify effective features, thus better segmentation results can be obtained. Finally, a multi-output combined loss function is proposed to make effective supervision for better segmentation. Experimental results demonstrate the promising performance of the proposed framework for chicken image segmentation.

Content-based human motion capture (MoCap) data retrieval facilitates reusing motion data that have already been captured and stored in a database. For a MoCap data retrieval system to get practically deployed, both high precision and natural interface are demanded. Targeting both, we propose a video-based human MoCap data retrieval solution in this work. It lets users to specify a query via a video clip, addresses the representational gap between video and MoCap clips and extracts discriminative motion features for precise retrieval. Specifically, the proposed scheme firstly converts each video clip or MoCap clip at a certain viewpoint to a binary silhouette sequence. Regarding a video or MoCap clip as a set of silhouette images, the proposed scheme uses a convolutional neural network, named MotionSet, to extract the discriminative motion feature of the clip. The extracted motion features are used to match a query to repository MoCap clips for the retrieval. Besides the algorithmic solution, we also contribute a human MoCap dataset and a human motion video dataset in couple that contain various action classes. Experiments show that our proposed scheme achieves an increase of around 0.25 in average MAP and costs about 1/26 time for online retrieval, when compared with the benchmark algorithm.

In order to address the realtime performance and robustness issues in UAV aerial image stitching, an improved SURF-based algorithm is proposed. This algorithm enhances the stitching process for UAV aerial images by first employing the Speeded-Up Robust Features SURF algorithm for rapid and robust feature point detection and localization. Subsequently, binary BRISK descriptors are used for feature description. In the feature matching phase, an initial coarse matching is performed using the Nearest-Neighbor Distance Ratio NNDR algorithm to filter out correct match pairs. Then, precise matching is achieved through the application of the PROSAC algorithm. Finally, a weighted fusion algorithm is employed for seamless image stitching. Experimental results demonstrate significant improvements in both matching efficiency and accuracy achieved by this algorithm. It serves as a robust and highly accurate solution with excellent stitching outcomes for UAV aerial image stitching.

In recent years, deep convolutional neural networks (DCNNs) have made strong progress on single-image super-resolution (SISR). Most of them work in a fully supervised way: they learn a non-linear mapping from a low-resolution (LR) image to its high-resolution (HR) version. The main problem is that this needs ground-truth HR images for supervision, which makes these methods hard to use widely. To ease this problem, we propose a weakly supervised framework named AASPNet (Adaptive Attention-Based Symmetric Progressive Network). It has two parts: an upsampling process and a downsampling process. Each part is built from several training units, but only the first unit needs supervision. In other words, with only $\times 2$ supervision, AASPNet can produce HR results at $\times 4$ and $\times 8$. The symmetry between the two parts is used in two ways. First, it gives consistency constraints that help recover more accurate HR images. Second, a cross self-supervision recovers HR images in a weakly supervised way, using both image and feature consistency between the upsampling and downsampling parts. We also add an adaptive attention strategy to pick out useful features, so the HR images get better. Experiments on several benchmark datasets show that our weakly supervised AASPNet reaches results that are close to state-of-the-art fully supervised DCNN methods.