1,721,194 research outputs found
Test-Time Synthetic-to-Real Adaptive Depth Estimation
No full textIs it possible for a synthetic to realistic domain adapted neural network in single image depth estimation to truly generalize on real world data? The resultant, adapted model will only generalize on the realistic domain dataset, which only reflects a small portion of the true, real world. As a result, the network still has to cope with the potential danger of domain shift between the realistic domain dataset and the real world data. Instead, a viable solution is to design the model to be capable of continuously adapting to the distribution of data it receives at test-time. In this paper, we propose a depth estimation method that is capable of adapting to the domain shift at test-time. Our method adapts to the unseen test-time domain, by updating the network using our proposed objective functions. Following former work, we reduce the entropy of the current prediction for refinement and adaptation. We propose a Logit Order Enforcement loss that can prevent the network from deviating into wrong solutions, which can result from the mere reduction of the aforementioned entropy. Qualitative and quantitative results show the effectiveness of our method. Our method reduces the dependency on training data by 5.8× on average, while achieving comparable performance to state-of-the-art unsupervised domain adaptation (UDA) and domain generalization methods (DG) on the KITTI dataset
SAR Image Generation of Ground Targets for Automatic Target Recognition Using Indirect Information
No full textThe effectiveness of using the simulated synthetic aperture radar (SAR) images of military targets in databases for automatic target recognition (ATR) is widely known. However, for simulated target images to be useful, they must be sufficiently similar to measured images; otherwise, they can degrade ATR performance. Two factors affect the quality of simulated SAR images: precision of the associated computer-aided design (CAD) model of the target and the accuracy and speed of the numerical techniques used to solve the electromagnetic problems in SAR image generation. In this study, a method for the 3D CAD modeling of the target is proposed; this method can be used when direct access to the target is not feasible and only indirect information is available. Further, a bistatic image formation concept based on the shooting-and-bouncing-ray technique is adopted; this concept helps achieve an accuracy comparable to that of the monostatic method. Moreover, it helps achieve a highly enhanced computation speed. In combination, these proposals provide an efficient and fast method to generate a database of simulated SAR images that can effectively support ATR activities. We demonstrate the effectiveness of the proposed method by comparing the simulated SAR images with the measured ones using structural similarity as a similarity measure; further, we evaluate the recognition rate obtained with the simulated images. We show that the used similarity measure bears a strong relation with the recognition rate, which is an aspect that may further contribute to considerable time savings when validating and refining simulated image databases.
Subspace-based Feature Alignment for Unsupervised Domain Adaptation
No full textAutonomous agents need to perceive the world in a robust way, such that the shift in data distribution does not lead to faulty perception results. When agents cannot be trained with abundant data, agents may need to operate on real world environments while trained on simulated data, and suffer from domain shift. This paper proposes an effective and robust unsupervised domain adaptation (UDA) method that can resolve these situations. In the UDA setup, we are given a labeled source domain and an unlabeled target domain that share the same set of classes but are sampled from different distributions. This domain shift prevents agents which employ deep neural networks from generalizing well on the target domain. Recent methods adopt the strategy of self-training the networks with pseudo labeled target samples. However, falsely labeled samples cause negative transfer and deteriorate generalization of a network. to reduce negative transfer we propose an algorithm that can filter the pseudo labels, and use the filtered labels to align the domains in the feature space. The samples whose labels have not passed the filtering process can be used as an index to tune the hyperparameters of our method. Across various benchmarks, we validate the performance of our method. Especially, our method achieves strong performance on the synthetic-to-real adaptation scenario
A Convex Relaxation of the Ambrosio–Tortorelli Elliptic Functionals for the Mumford–Shah Functional
No full textTexture Classification Based on Discriminative Component Selection of Local Binary Pattern and Variants
No full text
Differentiable Architecture Search Based on Coordinate Descent
No full textNeural architecture search (NAS) is an automated method searching for the optimal network architecture by optimizing the combinations of edges and operations. For efficiency, recent differentiable architecture search methods adopt a one-shot network, containing all the candidate operations in each edge, instead of sampling and training individual architectures. However, a recent study doubts the effectiveness of differentiable methods by showing that random search can achieve comparable performance with differentiable methods using the same search cost. Therefore, there is a need to reduce the search cost even for previous differentiable methods. For more efficient differentiable architecture search, we propose a differentiable architecture search based on coordinate descent (DARTS-CD) that searches for optimal operation over only one sampled edge per training step. DARTS-CD is proposed based on the coordinate descent algorithm, which is an efficient learning method for resolving large-scale problems by updating only a subset of parameters. In DARTS-CD, one edge is randomly sampled, in which all the operations are performed, whereas only one operation is applied to the other edges. Weight update is also performed only at the sampled edge. By optimizing each edge separately, as in the coordinate descent that optimizes each coordinate individually, DARTS-CD converges much faster than DARTS while using the network architecture similar to that used for evaluation. We experimentally show that DARTS-CD performs comparably to the state-of-the-art efficient architecture search algorithms, with an extremely low search cost of 0.125 GPU days (1/12 of the search cost of DARTS) on CIFAR-10 and CIFAR-100. Furthermore, a warm-up regularization method is introduced to improve the exploration capability, which further enhances the performance.
- …
