Unsupervised Domain Adaptation (UDA) can tackle the challenge that convolutional neural network(CNN)-based approaches for semantic segmentation heavily rely on the pixel-level annotated data, which is labor-intensive. However, existing UDA approaches in this regard inevitably require the full access to source datasets to reduce the gap between the source and target domains during model adaptation, which are impractical in the real scenarios where the source datasets are private, and thus cannot be released along with the well-trained source models. To cope with this issue, we propose a source-free domain adaptation framework for semantic segmentation, namely SFDA, in which only a well-trained source model and an unlabeled target domain dataset are available for adaptation. SFDA not only enables to recover and preserve the source domain knowledge from the source model via knowledge transfer during model adaptation, but also distills valuable information from the target domain for self-supervised learning. The pixel- and patch-level optimization objectives tailored for semantic segmentation are seamlessly integrated in the framework. The extensive experimental results on numerous benchmark datasets highlight the effectiveness of our framework against the existing UDA approaches relying on source data.
We present an architecture for information extraction from text that augments an existing parser with a character-level neural network. The network is trained using a measure of consistency of extracted data with existing databases as a form of noisy supervision. Our architecture combines the ability of constraint-based information extraction systems to easily incorporate domain knowledge and constraints with the ability of deep neural networks to leverage large amounts of data to learn complex features. Boosting the existing parser's precision, the system led to large improvements over a mature and highly tuned constraint-based production information extraction system used at Bloomberg for financial language text.
This paper introduces a visual-based localization method for autonomous vehicles (AVs) that operate in the absence of any complicated hardware system but a single camera. Visual localization refers to techniques that aim to find the location of an object based on visual information of its surrounding area. The problem of localization has been of interest for many years. However, visual localization is a relatively new subject in the literature of transportation. Moreover, the inevitable application of this type of localization in the context of autonomous vehicles demands special attention from the transportation community to this problem. This study proposes a two-step localization method that requires a database of geotagged images and a camera mounted on a vehicle that can take pictures while the car is moving. The first step which is image retrieval uses SIFT local feature descriptor to find an initial location for the vehicle using image matching. The next step is to utilize the Kalman filter to estimate a more accurate location for the vehicle as it is moving. All stages of the introduced method are implemented as a complete system using different Python libraries. The proposed system is tested on the KITTI dataset and has shown an average accuracy of 2 meters in finding the final location of the vehicle.
Statistical similarities between neuronal spike trains could reveal significant information on complex underlying processing. In general, the similarity between synchronous spike trains is somewhat easy to identify. However, the similar patterns also potentially appear in an asynchronous manner. However, existing methods for their identification tend to converge slowly, and cannot be applied to short sequences. In response, we propose Hellinger distance measure based on empirical probabilities, which we show to be as accurate as existing techniques, yet faster to converge for synthetic as well as experimental spike trains. Further, we cluster pairs of neuronal spike trains based on statistical similarities and found two non-overlapping classes, which could indicate functional similarities in neurons. Significantly, our technique detected functional heterogeneity in pairs of neuronal responses with the same performance as existing techniques, while exhibiting faster convergence. We expect the proposed method to facilitate large-scale studies of functional clustering, especially involving short sequences, which would in turn identify signatures of various diseases in terms of clustering patterns.
Segmentation of organs or lesions from medical images plays an essential role in many clinical applications such as diagnosis and treatment planning. Though Convolutional Neural Networks (CNN) have achieved the state-of-the-art performance for automatic segmentation, they are often limited by the lack of clinically acceptable accuracy and robustness in complex cases. Therefore, interactive segmentation is a practical alternative to these methods. However, traditional interactive segmentation methods require a large amount of user interactions, and recently proposed CNN-based interactive segmentation methods are limited by poor performance on previously unseen objects. To solve these problems, we propose a novel deep learning-based interactive segmentation method that not only has high efficiency due to only requiring clicks as user inputs but also generalizes well to a range of previously unseen objects. Specifically, we first encode user-provided interior margin points via our proposed exponentialized geodesic distance that enables a CNN to achieve a good initial segmentation result of both previously seen and unseen objects, then we use a novel information fusion method that combines the initial segmentation with only few additional user clicks to efficiently obtain a refined segmentation. We validated our proposed framework through extensive experiments on 2D and 3D medical image segmentation tasks with a wide range of previous unseen objects that were not present in the training set. Experimental results showed that our proposed framework 1) achieves accurate results with fewer user interactions and less time compared with state-of-the-art interactive frameworks and 2) generalizes well to previously unseen objects.
Fine-grained indoor localization has attracted attention recently because of the rapidly growing demand for indoor location-based services (ILBS). Specifically, massive (large-scale) multiple-input and multiple-output (MIMO) systems have received increasing attention due to high angular resolution. This paper presents an indoor localization testbed based on a massive MIMO orthogonal frequency-division multiplexing (OFDM) system, which supports physical-layer channel measurements. Instead of exploiting channel state information (CSI) directly for localization, we focus on positioning from the perspective of multipath components (MPCs), which are extracted from the CSI through the space-alternating generalized expectation-maximization (SAGE) algorithm. On top of the available MPCs, we propose a generalized fingerprinting system based on different single-metric and hybrid-metric schemes. We evaluate the impact of varying antenna topologies, feeding metrics, sizes of the training set, and fingerprinting methods. The experimental results show that the proposed fingerprinting method can achieve centimeter-level positioning accuracy with a relatively small training set. Specifically, the distributed uniform linear array obtains the highest accuracy with about 1.63-2.5-cm mean absolute errors resulting from the high spatial resolution.
Weakly-supervised Temporal Action Localization (WS-TAL) methods learn to localize temporal starts and ends of action instances in a video under only video-level supervision. Existing WS-TAL methods rely on deep features learned for action recognition. However, due to the mismatch between classification and localization, these features cannot distinguish the frequently co-occurring contextual background, i.e., the context, and the actual action instances. We term this challenge action-context confusion, and it will adversely affect the action localization accuracy. To address this challenge, we introduce a framework that learns two feature subspaces respectively for actions and their context. By explicitly accounting for action visual elements, the action instances can be localized more precisely without the distraction from the context. To facilitate the learning of these two feature subspaces with only video-level categorical labels, we leverage the predictions from both spatial and temporal streams for snippets grouping. In addition, an unsupervised learning task is introduced to make the proposed module focus on mining temporal information. The proposed approach outperforms state-of-the-art WS-TAL methods on three benchmarks, i.e., THUMOS14, ActivityNet v1.2 and v1.3 datasets.
The neuromorphic event cameras, which capture the optical changes of a scene, have drawn increasing attention due to their high speed and low power consumption. However, the event data are noisy, sparse, and nonuniform in the spatial-temporal domain with an extremely high temporal resolution, making it challenging to design backend algorithms for event-based vision. Existing methods encode events into point-cloud-based or voxel-based representations, but suffer from noise and/or information loss. Additionally, there is little research that systematically studies how to handle static and dynamic scenes with one universal design for event-based vision. This work proposes the Aligned Event Tensor (AET) as a novel event data representation, and a neat framework called Event Frame Net (EFN), which enables our model for event-based vision under static and dynamic scenes. The proposed AET and EFN are evaluated on various datasets, and proved to surpass existing state-of-the-art methods by large margins. Our method is also efficient and achieves the fastest inference speed among others.
In the framework of convolutional neural networks that lie at the heart of deep learning, downsampling is often performed with a max-pooling operation that only retains the element with maximum activation, while completely discarding the information contained in other elements in a pooling region. To address this issue, a novel pooling scheme, Ordinal Pooling Network (OPN), is introduced in this work. OPN rearranges all the elements of a pooling region in a sequence and assigns different weights to these elements based upon their orders in the sequence, where the weights are learned via the gradient-based optimisation. The results of our small-scale experiments on image classification task demonstrate that this scheme leads to a consistent improvement in the accuracy over max-pooling operation. This improvement is expected to increase in deeper networks, where several layers of pooling become necessary.
Image-only and pseudo-LiDAR representations are commonly used for monocular 3D object detection. However, methods based on them have shortcomings of either not well capturing the spatial relationships in neighbored image pixels or being hard to handle the noisy nature of the monocular pseudo-LiDAR point cloud. To overcome these issues, in this paper we propose a novel object-centric voxel representation tailored for monocular 3D object detection. Specifically, voxels are built on each object proposal, and their sizes are adaptively determined by the 3D spatial distribution of the points, allowing the noisy point cloud to be organized effectively within a voxel grid. This representation is proved to be able to locate the object in 3D space accurately. Furthermore, prior works would like to estimate the orientation via deep features extracted from an entire image or a noisy point cloud. By contrast, we argue that the local RoI information from the object image patch alone with a proper resizing scheme is a better input as it provides complete semantic clues meanwhile excludes irrelevant interferences. Besides, we decompose the confidence mechanism in monocular 3D object detection by considering the relationship between 3D objects and the associated 2D boxes. Evaluated on KITTI, our method outperforms state-of-the-art methods by a large margin. The code will be made publicly available soon.