Visual emotion recognition aims to associate images with appropriate emotions. There are different visual stimuli that can affect human emotion from low-level to high-level, such as color, texture, part, object, etc. However, most existing methods treat different levels of features as independent entity without having effective method for feature fusion. In this paper, we propose a unified CNN-RNN model to predict the emotion based on the fused features from different levels by exploiting the dependency among them. Our proposed architecture leverages convolutional neural network (CNN) with multiple layers to extract different levels of features with in a multi-task learning framework, in which two related loss functions are introduced to learn the feature representation. Considering the dependencies within the low-level and high-level features, a new bidirectional recurrent neural network (RNN) is proposed to integrate the learned features from different layers in the CNN model. Extensive experiments on both Internet images and art photo datasets demonstrate that our method outperforms the state-of-the-art methods with at least 7% performance improvement.

Autonomous driving is a crucial issue of the automobile industry, and research on lane change is its significant part. Previous works on the autonomous vehicle lane change mainly focused on lane change path planning and path tracking, but autonomous vehicle lane change decision making is rarely mentioned. Therefore, this paper establishes an autonomous lane change decision-making model based on benefit, safety, and tolerance by analyzing the factors of the autonomous vehicle lane change. Then, because of the multi-parameter and non-linearity of the autonomous lane change decision-making process, a support vector machine (SVM) algorithm with the Bayesian parameters optimization is adopted to solve this problem. Finally, we compare a lane change model based on rules with the proposed SVM model in the test set, and results illustrate that the SVM model performs better than the rule-based lane change model. Moreover, the real car experiment is carried out to verify the effectiveness of the decision model. INDEX TERMS Autonomous vehicle, lane change decision making, support vector machine, Bayesian optimization, drivers' habits.

This study investigated whether sound intensity affects listeners' sensitivity to a break in interaural correlation (BIC) embedded in wideband noise at different interaural delays. The results show that the detection duration threshold remained stable at the intensity between 60 and 70 dB SPL, but increased in accelerating fashion as the intensity decreased toward 40 dB SPL. Moreover, the threshold elevated linearly as the interaural delay increased from 0 to 4 ms, and the elevation slope became larger as the intensity decreased from 50 to 40 dB SPL. Thus, detecting the BIC is co-modulated by both intensity and interaural delay.

The subjective representation of the sounds delivered to the two ears of a human listener is closely associated with the interaural delay and correlation of these two-ear sounds. When the two-ear sounds, e.g., arbitrary noises, arrive simultaneously, the single auditory image of the binaurally identical noises becomes increasingly diffuse, and eventually separates into two auditory images as the interaural correlation decreases. When the interaural delay increases from zero to several milliseconds, the auditory image of the binaurally identical noises also changes from a single image to two distinct images. However, measuring the effect of these two factors on an identical group of participants has not been investigated. This study examined the impacts of interaural correlation and delay on detecting a binaurally uncorrelated fragment (interaural correlation = 0) embedded in the binaurally correlated noises (i.e., binaural gap or break in interaural correlation). We found that the minimum duration of the binaural gap for its detection (i.e., duration threshold) increased exponentially as the interaural delay between the binaurally identical noises increased linearly from 0 to 8 ms. When no interaural delay was introduced, the duration threshold also increased exponentially as the interaural correlation of the binaurally correlated noises decreased linearly from 1 to 0.4. A linear relationship between the effect of interaural delay and that of interaural correlation was described for listeners participating in this study: a 1 ms increase in interaural delay appeared to correspond to a 0.07 decrease in interaural correlation specific to raising the duration threshold. Our results imply that a tradeoff may exist between the impacts of interaural correlation and interaural delay on the subjective representation of sounds delivered to two human ears.