HiDAnet: RGB-D Salient Object Detection via Hierarchical Depth Awareness.

RGB-D saliency detection aims to fuse multi-modal cues to accurately localize salient regions. Existing works often adopt attention modules for feature modeling, with few methods explicitly leveraging fine-grained details to merge with semantic cues. Thus, despite the auxiliary depth information, it is still challenging for existing models to distinguish objects with similar appearances but at distinct camera distances. In this paper, from a new perspective, we propose a novel Hierarchical Depth Awareness network (HiDAnet) for RGB-D saliency detection. Our motivation comes from the observation that the multi-granularity properties of geometric priors correlate well with the neural network hierarchies. To realize multi-modal and multi-level fusion, we first use a granularity-based attention scheme to strengthen the discriminatory power of RGB and depth features separately. Then we introduce a unified cross dual-attention module for multi-modal and multi-level fusion in a coarse-to-fine manner. The encoded multi-modal features are gradually aggregated into a shared decoder. Further, we exploit a multi-scale loss to take full advantage of the hierarchical information. Extensive experiments on challenging benchmark datasets demonstrate that our HiDAnet performs favorably over the state-of-the-art methods by large margins. The source code can be found in https://github.com/Zongwei97/HIDANet/.

Reinforcement learning approach to control an inverted pendulum: A general framework for educational purposes.

Machine learning is often cited as a new paradigm in control theory, but is also often viewed as empirical and less intuitive for students than classical model-based methods. This is particularly the case for reinforcement learning, an approach that does not require any mathematical model to drive a system inside an unknown environment. This lack of intuition can be an obstacle to design experiments and implement this approach. Reversely there is a need to gain experience and intuition from experiments. In this article, we propose a general framework to reproduce successful experiments and simulations based on the inverted pendulum, a classic problem often used as a benchmark to evaluate control strategies. Two algorithms (basic Q-Learning and Deep Q-Networks (DQN)) are introduced, both in experiments and in simulation with a virtual environment, to give a comprehensive understanding of the approach and discuss its implementation on real systems. In experiments, we show that learning over a few hours is enough to control the pendulum with high accuracy. Simulations provide insights about the effect of each physical parameter and tests the feasibility and robustness of the approach.

OMNI-CONV: Generalization of the Omnidirectional Distortion-Aware Convolutions.

Omnidirectional images have drawn great research attention recently thanks to their great potential and performance in various computer vision tasks. However, processing such a type of image requires an adaptation to take into account spherical distortions. Therefore, it is not trivial to directly extend the conventional convolutional neural networks on omnidirectional images because CNNs were initially developed for perspective images. In this paper, we present a general method to adapt perspective convolutional networks to equirectangular images, forming a novel distortion-aware convolution. Our proposed solution can be regarded as a replacement for the existing convolutional network without requiring any additional training cost. To verify the generalization of our method, we conduct an analysis on three basic vision tasks, i.e., semantic segmentation, optical flow, and monocular depth. The experiments on both virtual and real outdoor scenarios show our adapted spherical models consistently outperform their counterparts.