Deep superpixel algorithms have made remarkable strides by substituting hand-crafted features with learnable ones. Nevertheless, we observe that existing deep superpixel methods, serving as mid-level representation operations, remain sensitive to the statistical properties (e.g., color distribution, high-level semantics) embedded within the training dataset. Consequently, learnable features exhibit constrained discriminative capability, resulting in unsatisfactory pixel grouping performance, particularly in untrainable application scenarios. To address this issue, we propose the Content Disentangle Superpixel (CDS) algorithm to selectively separate the invariant inter-pixel correlations and statistical properties, i.e., style noise. Specifically, We first construct auxiliary modalities that are homologous to the original RGB image but have substantial stylistic variations. Then, driven by mutual information, we propose the local-grid correlation alignment across modalities to reduce the distribution discrepancy of adaptively selected features and learn invariant inter-pixel correlations. Afterwards, we perform global-style mutual information minimization to enforce the separation of invariant content and train data styles. The experimental results on four benchmark datasets demonstrate the superiority of our approach to existing state-of-the-art methods, regarding boundary adherence, generalization, and efficiency. Code and pre-trained model are available at https://github.com/rookiie/CDSpixel.

Audio-visual segmentation (AVS) aims to segment the object instances that produce sound at the time of the video frames. Existing related solutions focus on designing cross-modal interaction mechanisms, which try to learn audio-visual correlations and simultaneously segment objects. Despite effectiveness, the close-coupling network structures become increasingly complex and hard to analyze. To address these problems, we propose a simple but effective method, ‘EachPerformsItsFunctions (PIF),’ which focuses on task decomposition and feature assignment. Inspired by human sensory experiences, PIF decouples AVS into two subtasks, correlation learning, and segmentation refinement, via two branches. Correlation learning aims to learn the correspondence between sound and visible individuals and provide the positional prior. Segmentation refinement focuses on fine segmentation. Then we assign different level features to perform the appropriate duties, i.e., using deep features for cross-modal interaction due to their semantic advantages; using rich textures of shallow features to improve segmentation results. Moreover, we propose the recurrent collaboration block to enhance interbranch communication. Experimental results on AVSBench show that our method outperforms related state-of-the-art methods by a large margin (e.g., +6.0% mIoU and +7.6% F-score on the Multi-Source subset). In addition, by purposely boosting subtasks' performance, our approach can serve as a strong baseline for audio-visual segmentation.

Barrel distortion rectification aims at removing the radial distortion in a distorted image captured by a wide-angle lens. Previous deep learning methods mainly solve this problem by learning the implicit distortion parameters or the nonlinear rectified mapping function in a direct manner. However, this type of manner results in an indistinct learning process of rectification and thus limits the deep perception of distortion. In this paper, inspired by the curriculum learning, we analyze the barrel distortion rectification task in a progressive and meaningful manner. By considering the relationship among different construction levels in an image, we design a multi-level curriculum that disassembles the rectification task into three levels, structure recovery, semantics embedding, and texture rendering. With the guidance of the curriculum that corresponds to the construction of images, the proposed hierarchical architecture enables a progressive rectification and achieves more accurate results. Moreover, we present a novel distortion-aware pre-training strategy to facilitate the initial learning of neural networks, promoting the model to converge faster and better. Experimental results on the synthesized and real-world distorted image datasets show that the proposed approach significantly outperforms other learning methods, both qualitatively and quantitatively.

Convolutional neural network (CNN) is the primary technique that has greatly promoted the development of computer vision technologies. However, there is little research on how to allocate parameters in different convolution layers when designing CNNs. We research mainly on revealing the relationship between CNN parameter distribution, i.e., the allocation of parameters in convolution layers, and the discriminative performance of CNN. Unlike previous works, we do not append more elements into the network, such as more convolution layers or denser short connections. We focus on enhancing the discriminative performance of CNN through varying its parameter distribution under strict size constraint. We propose an energy function to represent the CNN parameter distribution, which establishes the connection between the allocation of parameters and the discriminative performance of CNN. Extensive experiments with shallow CNNs on three public image classification data sets demonstrate that the CNN parameter distribution with a higher energy value will promote the model to obtain better performance. According to the motivated observation, the problem of finding the optimal parameter distribution can be transformed into an optimization problem of finding the biggest energy value. We present a simple yet effective guideline that uses balanced parameter distribution to design CNNs. Extensive experiments on ImageNet with three popular backbones, i.e., AlexNet, ResNet34, and ResNet101, demonstrate that the proposed guideline can make consistent improvements upon different baselines under strict size constraint.

In content-based image retrieval (CBIR), one of the most challenging and ambiguous tasks are to correctly understand the human query intention and measure its semantic relevance with images in the database. Due to the impressive capability of visual saliency in predicting human visual attention that is closely related to the query intention, this paper attempts to explicitly discover the essential effect of visual saliency in CBIR via qualitative and quantitative experiments. Toward this end, we first generate the fixation density maps of images from a widely used CBIR dataset by using an eye-tracking apparatus. These ground-truth saliency maps are then used to measure the influence of visual saliency to the task of CBIR by exploring several probable ways of incorporating such saliency cues into the retrieval process. We find that visual saliency is indeed beneficial to the CBIR task, and the best saliency involving scheme is possibly different for different image retrieval models. Inspired by the findings, this paper presents two-stream attentive CNNs with saliency embedded inside for CBIR. The proposed network has two streams that simultaneously handle two tasks. The main stream focuses on extracting discriminative visual features that are tightly related to semantic attributes. Meanwhile, the auxiliary stream aims to facilitate the main stream by redirecting the feature extraction to the salient image content that human may pay attention to. By fusing these two streams into the Main and Auxiliary CNNs (MAC), image similarity can be computed as the human being does by reserving conspicuous content and suppressing irrelevant regions. Extensive experiments show that the proposed model achieves impressive performance in image retrieval on four public datasets.

Nowadays cross-media retrieval is an useful technology that helps people find expected information from the huge amount of multimodal data more efficiently. A common cross-media retrieval framework is first to map features of different modalities into an isomorphic semantic space so that the similarity between heterogeneous data can be measured. For most of semantic space based methods, the mapping mechanism from original to semantic space of each modality is optimized independently, yet the more discriminative characteristic of a certain modality is not taken into account. In this paper, we propose a deep framework which introduces a latent embedding layer to learn joint parameters to obtain semantically meaningful representations of images and texts. Specifically, the discriminative characteristic embedded in the textual modality can be transferred to images through the latent embedding layer and joint parameters to enhance the consistency between semantic representations. Extensive experiments on the three popular publicly available datasets well demonstrate the superiority of the proposed method, which achieves the new state-of-the-arts.

Although deep convolutional neural networks (CNNs) have significantly boosted the performance of many computer vision tasks, their complexities~(the size or the number of parameters) are also dramatically increased even with slight performance improvement. However, the larger network leads to more computation requirements, which are unfavorable to resource-constrained scenarios, such as the widely used embedded systems. In this paper, we tentatively explore the essential effect of CNN parameter layout, ıe, the allocation of parameters in the convolution layers, on the discriminative capability of CNN. Instead of enlarging the breadth or depth of networks, we attempt to improve the discriminative ability of CNN by changing its parameter layout under strict size constraint. Toward this end, a novel energy function is proposed to represent the CNN parameter layout, which makes it possible to model the relationship between the allocation of parameters in the convolution layers and the discriminative ability of CNN. According to extensive experimental results with plain CNN models and Residual Nets, we find that the higher the energy of a specific CNN parameter layout is, the better its discriminative ability is. Following this finding, we propose a novel approach to learn the better parameter layout. Experimental results on two public image classification datasets show that the CNN models with the learned parameter layouts achieve the better image classification results under strict size constraint.

Generally distance-based similarity estimation between two images is not always reliable due to the limitations in both image understanding techniques and distance measure methods. This paper presents a novel approach for improving the similarity estimation through introducing the distribution information of similarity scores. The key idea is based on an underlying assumption that the distributions of similarity scores are similar for true-relevant images when they query an independent database. By representing each distribution with the area under the corresponding similarity score curve, the difference between different distributions can be easily calculated and employed to update the original distance measure. Experiments on three public datasets with various feature representations show that the enhanced similarity estimation remarkably outperforms the original distance measure and the proposed approach also keeps a good generalization ability on various datasets and feature representations.