In autonomous driving, it is crucial to train a single segmentation model that can generalize well on various target environments. Due to the lack of pixel-level annotation and a large domain discrepancy between domain pairs, it could be tough to achieve encouraging performance for multi-target domain adaptive semantic segmentation. To this end, we propose a novel Multi-level Collaborative Learning (MCL) framework that consists of two core components, namely Multi-level Self-Training (MST) and Hierarchical Knowledge Distillation (HKD). Specifically, MST focuses on individual, collaborative, and ensemble learning, whilst HKD aims to play the model’s ensemble capability. These designs enable the proposed MCL to fully exploit the multiple target data to train more powerful teachers and yield more accurate domain alignment. In addition, we integrate style transfer, self-training, and knowledge distillation into an end-to-end training scheme, making the proposed MCL more practical in applications. Empirically, we conduct extensive experiments on multi-target benchmarks. The encouraging results show the effectiveness of our method and state-of-the-art performance has been achieved. Codes are available athttps://github.com/feifei-cv/MCL.

Unsupervised domain adaptation (UDA) attempts to learn domain invariant representations and has achieved significant progress, whereas self-training-based UDA methods have shown powerful performance. However, due to the domain gap, pseudo-labels selected through high confidence scores or uncertainty inevitably contain noise, leading to inaccurate predictions. To address this issue, we propose a novel risk-consistent training method. Specifically, both clean and noisy classifiers are introduced to estimate the noise transition matrix. The clean classifier is exploited to assign pseudo-labels for target data in each iteration. The noisy classifier is then trained with noisy target samples, and the optimal parameters are obtained through a closed-form solution. Heuristically, we also pre-train a domain predictor to select a target-like source example for the noise transition matrix estimation. In addition, we design an uncertainty-guided regularization to generate soft pseudo-labels and avoid overconfident predictions. Extensive experimental results show the effectiveness of our method, and state-of-the-art performance has been achieved. Codes are available athttps://github.com/feifei-cv/RCE.

The range resolution of high-resolution wideband radar is much smaller than the target size. Its echo signals tend to be diverse and sensitive to small changes of targets. Therefore, it is difficult to capture and distinguish the features in radar signals. In this article, we propose a radar target recognition pipeline based on a deep nested neural network. The framework consists of three parts: The translation sensitivity of the training data is first addressed in the preprocessing section. The second step is to obtain an embedded representation of the radar echo signals by the combination of the adjustment layer, convolutional neural network (CNN), and the squeeze and excitation (SE) block. Finally, the target is recognized through inputting embedded representation as a time sequence into the stacked bidirectional recurrent neural network (bi-RNN) based on an attention mechanism. Compared with the traditional methods, the proposed deep nested neural network extracts and takes advantage of the features of radar echo signals more effectively, including the envelope features and local physical structural features. The experimental results based on the test data indicate that the proposed method has a great advantage over other methods in the case of large data sets as well as small training data sets and is robust to the small translation of test samples and noises, exhibiting high engineering practical value.

Visual field defects (VFD) can be caused by a variety of conditions. Checking and tracking the progression of VFD is an important part of an eye assessment. Although the use of standard automatic perimetry (SAP) is very popular for VFD diagnosis, it limits the population because of its high requirement for patients. We used a video-watching task as a replacement modality, which precludes the long period of fixation and uses the on-screen gaze to replace the button response. We developed a simulation system to mimic the different types of VFD in people with a normal pattern.We hypothesize that patients with VFD need more eye movement to compensate for the unseen area. We proposed a metric that indicates the gross eye movements toward a specific direction and found a significant difference between the VFD and normal pattern. Furthermore, we found videos that show the unique eye movement pattern in different eye conditions.

An important application of affective image annotation is affective image content analysis, which aims to automatically understand the emotion being brought to viewers by image contents. The so-called subjective perception issue, i.e., different viewers may have different emotional responses to the same image, makes it difficult to link image features with the expected perceived emotion. Due to the ability to learn features, recent deep learning technologies have opened a new window on affective image content analysis, which has led to a growing demand for affective image annotation technologies to build large reliable training datasets. This paper proposes a novel affective image annotation technique, AffectI, for efficiently collecting diverse and reliable emotional labels with the estimate emotion distribution for images based on the concept of Game With a Purpose (GWAP). AffectI features three novel mechanisms: a selection mechanism for ensuring all emotion words being fairly evaluated for collecting diverse and reliable labels; an estimation mechanism for estimating the emotion distribution by aggregating partial pairwise comparisons of the emotion words for collecting the labels effectively and efficiently; an incentive mechanism shows the comparison between current player and her opponents as well as all past players to promote the interest of players and also contributes the reliability and diversity. Our experimental results demonstrate that AffectI is superior to existing methods in terms of being able to collect more diverse and reliable labels. The advantage of using GWAP for reducing the frustration of evaluators was also confirmed through subjective evaluation.

Accurately predicting video saliency is important for applications such as video quality assessment, summary, compression, and retargeting. As the automatic saliency models for videos suffer from problems of inaccuracy, determining video saliency from data on the human gaze is a promising approach. Due to differences in individual observers, however, eye-tracking data of a certain number of observers are usually required to compute a visual attention map close to the ground truth. Although it has become cheaper to acquire human eye-tracking data thanks to the lower price of equipment, it is still not easy to carry out studies with a large number of observers. To keep the balance between accuracy and expense, this paper proposes a new method for suggesting the appropriate number of observers needed in eye-tracking experiments for a given video. Through carefully analyzing eye-tracking data of various video clips, we found videos can be classified into four types based on the number of observers required to approach the ground truth. A new support vector machine (SVM) classifier was trained to automatically classify videos into one of the four typical types.

With the rapid development of Internet technology, images on the Internet are used in various aspects of people’s lives. The security and authorization of images are strongly dependent on image quality. Some potential problems have also emerged, among which the quality assessment and denoising of images are particularly evident. This paper proposes a novel NR-IQA method based on the dual convolutional neural network structure, which combines saliency detection with the human visual system (HSV), used as a weighting function to reflect the important distortion caused by the local area. The model is trained using gray and color features in the HSV space. It is applied to the parameter selection of an image denoising algorithm. The experiment proves that our proposed method can accurately evaluate image quality in the process of denoising. It provides great help in parameter optimization iteration and improves the performance of the algorithm. Through experiments, we obtain both improved image quality and a reasonable result of subject assessment when the cascaded algorithm is applied in image security and authorization.