Attention mechanisms are widely used in current encoder/decoder frameworks of image captioning, where a weighted average on encoded vectors is generated at each time step to guide the caption decoding process. However, the decoder has little idea of whether or how well the attended vector and the given attention query are related, which could make the decoder give misled results. In this paper, we propose an Attention on Attention (AoA) module, which extends the conventional attention mechanisms to determine the relevance between attention results and queries. AoA first generates an information vector and an attention gate using the attention result and the current context, then adds another attention by applying element-wise multiplication to them and finally obtains the attended information, the expected useful knowledge. We apply AoA to both the encoder and the decoder of our image captioning model, which we name as AoA Network (AoANet). Experiments show that AoANet outperforms all previously published methods and achieves a new state-of-the-art performance of 129.8 CIDEr-D score on MS COCO Karpathy offline test split and 129.6 CIDEr-D (C40) score on the official online testing server. Code is available at https://github.com/husthuaan/AoANet.

Low-light images are not conducive to human observation and computer vision algorithms due to their low visibility. To solve this problem, many image enhancement techniques have been proposed. However, existing techniques inevitably introduce color and lightness distortion when increasing visibility. To lower the distortion, we propose a novel enhancement method using the response characteristics of cameras. First, we investigate the relationship between two images with different exposures to obtain an accurate camera response model. Then we borrow the illumination estimation techniques to estimate the exposure ratio map. Finally, we use our camera response model to adjust each pixel to its desired exposure according to the estimated exposure ratio map. Experiments show that our method can obtain enhancement results with less color and lightness distortion compared to several state-of-the-art methods.

In this paper, a novel united low-light image enhancement framework for both contrast enhancement and denoising is proposed. First, the low-light image is segmented into superpixels, and the ratio between the local standard deviation and the local gradients is utilized to estimate the noise-texture level of each superpixel. Then the image is inverted to be processed in the following steps. Based on the noise-texture level, a smooth base layer is adaptively extracted by the BM3D filter, and another detail layer is extracted by the first order differential of the inverted image and smoothed with the structural filter. These two layers are adaptively combined to get a noise-free and detail-preserved image. At last, an adaptive enhancement parameter is adopt into the dark channel prior dehazing process to enlarge contrast and prevent over/under enhancement. Experimental results demonstrate that our proposed method outperforms traditional methods in both subjective and objective assessments.

Deep learning, a crucial technique for achieving artificial intelligence (AI), has been successfully applied in many fields. The gradual application of the latest architectures of deep learning in the field of time series forecasting (TSF), such as Transformers, has shown excellent performance and results compared to traditional statistical methods. These applications are widely present in academia and in our daily lives, covering many areas including forecasting electricity consumption in power systems, meteorological rainfall, traffic flow, quantitative trading, risk control in finance, sales operations and price predictions for commercial companies, and pandemic prediction in the medical field. Deep learning-based TSF tasks stand out as one of the most valuable AI scenarios for research, playing an important role in explaining complex real-world phenomena. However, deep learning models still face challenges: they need to deal with the challenge of large-scale data in the information age, achieve longer forecasting ranges, reduce excessively high computational complexity, etc. Therefore, novel methods and more effective solutions are essential. In this paper, we review the latest developments in deep learning for TSF. We begin by introducing the recent development trends in the field of TSF and then propose a new taxonomy from the perspective of deep neural network models, comprehensively covering articles published over the past five years. We also organize commonly used experimental evaluation metrics and datasets. Finally, we point out current issues with the existing solutions and suggest promising future directions in the field of deep learning combined with TSF. This paper is the most comprehensive review related to TSF in recent years and will provide a detailed index for researchers in this field and those who are just starting out.

Saliency detection aims to detect the most attractive objects in images, which has been widely used as a foundation for various multimedia applications. In this paper, we propose a novel salient object detection algorithm for RGB-D images using center-dark channel prior. First, we generate an initial saliency map based on a color saliency map and a depth saliency map of a given RGB-D image. Then, we generate a center-dark channel map based on a center saliency prior and a dark channel prior. Finally, we fuse the initial saliency map with the center dark channel map to generate the final saliency map. The proposed algorithm is evaluated on two public RGB-D datasets, and the experimental results show that our method outperforms the state-of-the-art methods.