In this paper, we make the first attempt to build a framework to simultaneously estimate semantic parts, shape, translation, and orientation of cars from single street view. Our framework contains three major contributions. Firstly, a novel domain adaptation approach based on the class consistency loss is developed to transfer our part segmentation model from the synthesized images to the real images. Secondly, we propose a novel network structure that leverages part-level features from street views and 3D losses for pose and shape estimation. Thirdly, we construct a high quality dataset that contains more than 300 different car models with physical dimensions and part-level annotations based on global and local deformations. We have conducted experiments on both synthesized data and real images. Our results show that the domain adaptation approach can bring 35.5 percentage point performance improvement in terms of mean intersection-over-union score (mIoU) comparing with the baseline network using domain randomization only. Our network for translation and orientation estimation achieves competitive performance on highly complex street views (e.g., 11 cars per image on average). Moreover, our network is able to reconstruct a list of 3D car models with part-level details from street views, which could benefit various applications such as fine-grained car recognition, vehicle re-identification, and traffic simulation.
Due to the sparse rewards and high degree of environment variation, reinforcement learning approaches such as Deep Deterministic Policy Gradient (DDPG) are plagued by issues of high variance when applied in complex real world environments. We present a new framework for overcoming these issues by incorporating a stochastic switch, allowing an agent to choose between high and low variance policies. The stochastic switch can be jointly trained with the original DDPG in the same framework. In this paper, we demonstrate the power of the framework in a navigation task, where the robot can dynamically choose to learn through exploration, or to use the output of a heuristic controller as guidance. Instead of starting from completely random moves, the navigation capability of a robot can be quickly bootstrapped by several simple independent controllers. The experimental results show that with the aid of stochastic guidance we are able to effectively and efficiently train DDPG navigation policies and achieve significantly better performance than state-of-the-art baselines models.
Semi-supervised learning is crucial for alleviating labelling burdens in people-centric sensing. However, human-generated data inherently suffer from distribution shift in semi-supervised learning due to the diverse biological conditions and behavior patterns of humans. To address this problem, we propose a generic distributionally robust model for semi-supervised learning on distributionally shifted data. Considering both the discrepancy and the consistency between the labeled data and the unlabeled data, we learn the latent features that reduce person-specific discrepancy and preserve task-specific consistency. We evaluate our model in a variety of people-centric recognition tasks on real-world datasets, including intention recognition, activity recognition, muscular movement recognition and gesture recognition. The experiment results demonstrate that the proposed model outperforms the state-of-the-art methods.
The ability to interact and understand the environment is a fundamental prerequisite for a wide range of applications from robotics to augmented reality. In particular, predicting how deformable objects will react to applied forces in real time is a significant challenge. This is further confounded by the fact that shape information about encountered objects in the real world is often impaired by occlusions, noise and missing regions e.g. a robot manipulating an object will only be able to observe a partial view of the entire solid. In this work we present a framework, 3D-PhysNet, which is able to predict how a three-dimensional solid will deform under an applied force using intuitive physics modelling. In particular, we propose a new method to encode the physical properties of the material and the applied force, enabling generalisation over materials. The key is to combine deep variational autoencoders with adversarial training, conditioned on the applied force and the material properties. We further propose a cascaded architecture that takes a single 2.5D depth view of the object and predicts its deformation. Training data is provided by a physics simulator. The network is fast enough to be used in real-time applications from partial views. Experimental results show the viability and the generalisation properties of the proposed architecture.
We study the problem of recovering an underlying 3D shape from a set of images. Existing learning based approaches usually resort to recurrent neural nets, e.g., GRU, or intuitive pooling operations, e.g., max/mean pooling, to fuse multiple deep features encoded from input images. However, GRU based approaches are unable to consistently estimate 3D shapes given the same set of input images as the recurrent unit is permutation variant. It is also unlikely to refine the 3D shape given more images due to the long-term memory loss of GRU. The widely used pooling approaches are limited to capturing only the first order/moment information, ignoring other valuable features. In this paper, we present a new feed-forward neural module, named AttSets, together with a dedicated training algorithm, named JTSO, to attentionally aggregate an arbitrary sized deep feature set for multi-view 3D reconstruction. AttSets is permutation invariant, computationally efficient, flexible and robust to multiple input images. We thoroughly evaluate various properties of AttSets on large public datasets. Extensive experiments show AttSets together with JTSO algorithm significantly outperforms existing aggregation approaches.
Many natural language processing tasks solely rely on sparse dependencies between a few tokens in a sentence. Soft attention mechanisms show promising performance in modeling local/global dependencies by soft probabilities between every two tokens, but they are not effective and efficient when applied to long sentences. By contrast, hard attention mechanisms directly select a subset of tokens but are difficult and inefficient to train due to their combinatorial nature. In this paper, we integrate both soft and hard attention into one context fusion model, "reinforced self-attention (ReSA)", for the mutual benefit of each other. In ReSA, a hard attention trims a sequence for a soft self-attention to process, while the soft attention feeds reward signals back to facilitate the training of the hard one. For this purpose, we develop a novel hard attention called "reinforced sequence sampling (RSS)", selecting tokens in parallel and trained via policy gradient. Using two RSS modules, ReSA efficiently extracts the sparse dependencies between each pair of selected tokens. We finally propose an RNN/CNN-free sentence-encoding model, "reinforced self-attention network (ReSAN)", solely based on ReSA. It achieves state-of-the-art performance on both Stanford Natural Language Inference (SNLI) and Sentences Involving Compositional Knowledge (SICK) datasets.
Multimodal wearable sensor data classification plays an important role in ubiquitous computing and has a wide range of applications in scenarios from healthcare to entertainment. However, most existing work in this field employs domain-specific approaches and is thus ineffective in complex sit- uations where multi-modality sensor data are col- lected. Moreover, the wearable sensor data are less informative than the conventional data such as texts or images. In this paper, to improve the adapt- ability of such classification methods across differ- ent application domains, we turn this classification task into a game and apply a deep reinforcement learning scheme to deal with complex situations dynamically. Additionally, we introduce a selective attention mechanism into the reinforcement learn- ing scheme to focus on the crucial dimensions of the data. This mechanism helps to capture extra information from the signal and thus it is able to significantly improve the discriminative power of the classifier. We carry out several experiments on three wearable sensor datasets and demonstrate the competitive performance of the proposed approach compared to several state-of-the-art baselines.
Modelling the physical properties of everyday objects is a fundamental prerequisite for autonomous robots. We present a novel generative adversarial network (Defo-Net), able to predict body deformations under external forces from a single RGB-D image. The network is based on an invertible conditional Generative Adversarial Network (IcGAN) and is trained on a collection of different objects of interest generated by a physical finite element model simulator. Defo-Net inherits the generalisation properties of GANs. This means that the network is able to reconstruct the whole 3-D appearance of the object given a single depth view of the object and to generalise to unseen object configurations. Contrary to traditional finite element methods, our approach is fast enough to be used in real-time applications. We apply the network to the problem of safe and fast navigation of mobile robots carrying payloads over different obstacles and floor materials. Experimental results in real scenarios show how a robot equipped with an RGB-D camera can use the network to predict terrain deformations under different payload configurations and use this to avoid unsafe areas.
For vehicle autonomy, driver assistance and situational awareness, it is necessary to operate at day and night, and in all weather conditions. In particular, long wave infrared (LWIR) sensors that receive predominantly emitted radiation have the capability to operate at night as well as during the day. In this work, we employ a polarised LWIR (POL-LWIR) camera to acquire data from a mobile vehicle, to compare and contrast four different convolutional neural network (CNN) configurations to detect other vehicles in video sequences. We evaluate two distinct and promising approaches, two-stage detection (Faster-RCNN) and one-stage detection (SSD), in four different configurations. We also employ two different image decompositions: the first based on the polarisation ellipse and the second on the Stokes parameters themselves. To evaluate our approach, the experimental trials were quantified by mean average precision (mAP) and processing time, showing a clear trade-off between the two factors. For example, the best mAP result of 80.94% was achieved using Faster-RCNN, but at a frame rate of 6.4 fps. In contrast, MobileNet SSD achieved only 64.51% mAP, but at 53.4 fps.
We propose a novel monocular visual odometry (VO) system called UnDeepVO in this paper. UnDeepVO is able to estimate the 6-DoF pose of a monocular camera and the depth of its view by using deep neural networks. There are two salient features of the proposed UnDeepVO: one is the unsupervised deep learning scheme, and the other is the absolute scale recovery. Specifically, we train UnDeepVO by using stereo image pairs to recover the scale but test it by using consecutive monocular images. Thus, UnDeepVO is a monocular system. The loss function defined for training the networks is based on spatial and temporal dense information. A system overview is shown in Fig. 1. The experiments on KITTI dataset show our UnDeepVO achieves good performance in terms of pose accuracy.