When building a unified vision system or gradually adding new capabilities to a system, the usual assumption is that training data for all tasks is always available. However, as the number of tasks grows, storing and retraining on such data becomes infeasible. A new problem arises where we add new capabilities to a Convolutional Neural Network (CNN), but the training data for its existing capabilities are unavailable. We propose our Learning without Forgetting method, which uses only new task data to train the network while preserving the original capabilities. Our method performs favorably compared to commonly used feature extraction and fine-tuning adaption techniques and performs similarly to multitask learning that uses original task data we assume unavailable. A more surprising observation is that Learning without Forgetting may be able to replace fine-tuning with similar old and new task datasets for improved new task performance.

Learning without Forgetting

Point cloud learning has lately attracted increasing attention due to its wide applications in many areas, such as computer vision, autonomous driving, and robotics As a dominating technique in AI, deep learning has been successfully used to solve various 2D vision problems However, deep learning on point clouds is still in its infancy due to the unique challenges faced by the processing of point clouds with deep neural networks Recently, deep learning on point clouds has become even thriving, with numerous methods being proposed to address different problems in this area To stimulate future research, this paper presents a comprehensive review of recent progress in deep learning methods for point clouds It covers three major tasks, including 3D shape classification, 3D object detection and tracking, and 3D point cloud segmentation It also presents comparative results on several publicly available datasets, together with insightful observations and inspiring future research directions

Deep Learning for 3D Point Clouds: A Survey

Image Super-Resolution (SR) is an important class of image processing techniqueso enhance the resolution of images and videos in computer vision. Recent years have witnessed remarkable progress of image super-resolution using deep learning techniques. This article aims to provide a comprehensive survey on recent advances of image super-resolution using deep learning approaches. In general, we can roughly group the existing studies of SR techniques into three major categories: supervised SR, unsupervised SR, and domain-specific SR. In addition, we also cover some other important issues, such as publicly available benchmark datasets and performance evaluation metrics. Finally, we conclude this survey by highlighting several future directions and open issues which should be further addressed by the community in the future.

Deep Learning for Image Super-Resolution: A Survey

European conference on computer vision (ECCV)

Stereo image pairs can be used to improve the performance of super-resolution (SR) since additional information is provided from a second viewpoint. However, it is challenging to incorporate this information for SR since disparities between stereo images vary significantly. In this paper, we propose a parallax-attention stereo superresolution network (PASSRnet) to integrate the information from a stereo image pair for SR. Specifically, we introduce a parallax-attention mechanism with a global receptive field along the epipolar line to handle different stereo images with large disparity variations. We also propose a new and the largest dataset for stereo image SR (namely, Flickr1024). Extensive experiments demonstrate that the parallax-attention mechanism can capture correspondence between stereo images to improve SR performance with a small computational and memory cost. Comparative results show that our PASSRnet achieves the state-of-the-art performance on the Middlebury, KITTI 2012 and KITTI 2015 datasets.

/pdf/learning-parallax-attention-for-stereo-image-super-1zf1s5eej0.pdf

Learning Parallax Attention for Stereo Image Super-Resolution

Most existing CNN-based super-resolution (SR) methods are developed based on an assumption that the degradation is fixed and known (e.g., bicubic downsampling). However, these methods suffer a severe performance drop when the real degradation is different from their assumption. To handle various unknown degradations in real-world applications, previous methods rely on degradation estimation to reconstruct the SR image. Nevertheless, degradation estimation methods are usually time-consuming and may lead to SR failure due to large estimation errors. In this paper, we propose an unsupervised degradation representation learning scheme for blind SR without explicit degradation estimation. Specifically, we learn abstract representations to distinguish various degradations in the representation space rather than explicit estimation in the pixel space. Moreover, we introduce a Degradation-Aware SR (DASR) network with flexible adaption to various degradations based on the learned representations. It is demonstrated that our degradation representation learning scheme can extract discriminative representations to obtain accurate degradation information. Experiments on both synthetic and real images show that our network achieves state-of-the-art performance for the blind SR task. Code is available at: https://github.com/LongguangWang/DASR.

/pdf/unsupervised-degradation-representation-learning-for-blind-2nsktuxwxq.pdf

Unsupervised Degradation Representation Learning for Blind Super-Resolution

Shape representation for 3-D models is an important topic in computer vision, multimedia analysis, and computer graphics. Recent multiview-based methods demonstrate promising performance for 3-D shape recognition and retrieval. However, most multiview-based methods ignore the correlations of multiple views or suffer from high computional cost. In this paper, we propose a novel multiview-based network architecture for 3-D shape recognition and retrieval. Our network combines convolutional neural networks (CNNs) with long short-term memory (LSTM) to exploit the correlative information from multiple views. Well-pretrained CNNs with residual connections are first used to extract a low-level feature of each view image rendered from a 3-D shape. Then, a LSTM and a sequence voting layer are employed to aggregate these features into a shape descriptor. The highway network and a three-step training strategy are also adopted to boost the optimization of the deep network. Experimental results on two public datasets demonstrate that the proposed method achieves promising performance for 3-D shape recognition and the state-of-the-art performance for the 3-D shape retrieval.

Learning Multi-View Representation With LSTM for 3-D Shape Recognition and Retrieval

Stepped-frequency waveforms can achieve an ultrawide bandwidth by using a sequence of single-frequency pulses. The advantages of stepped-frequency waveforms are low hardware requirements and high resolution. However, the stepped-frequency waveform requires a long time period to transmit the signals, which limits its application in synthetic aperture radar (SAR). The available imaging range width is usually very narrow, unless the range and azimuth resolutions are both decreased. In this paper, a random-frequency SAR imaging scheme based on compressed sensing is proposed. If the targets are sparse or compressible, it is sufficient to transmit only a small number of random frequencies to reconstruct the image of the targets. This means that the limitations of the stepped-frequency technique for SAR can be overcome. The available imaging range width can be enlarged significantly, while the range and azimuth resolutions are both maintained. Random undersampling is very easy to implement for both range and azimuth dimensions, and no new hardware components are needed. Simulation and experimental results are presented to demonstrate the validity of the proposed method.

Random-Frequency SAR Imaging Based on Compressed Sensing

Robust detection of infrared small and dim targets with highly heterogeneous backgrounds plays an indispensable role in infrared search and tracking (IRST) system, which is still a challenging problem. In this article, a novel method based on multisubspace learning and spatial-temporal tensor data structure is aimed to solve this problem. First, a tensor data structure is constructed to use inner correlation in spatial and temporal domain of the infrared image sequence. Second, in consideration of the complex and heterogeneous backgrounds in infrared images, the proposed method promotes the multisubspace property to tensor domain to separate the target and the background more accurately. Finally, an efficient and effective optimization algorithm based on alternating direction method of multipliers (ADMM) is designed to solve this problem. Experimental results on various and real scenes demonstrate the superiority of the proposed method compared to other five baseline methods.

Jungang Yang

Papers

Learning Parallax Attention for Stereo Image Super-Resolution

Unsupervised Degradation Representation Learning for Blind Super-Resolution

Learning Multi-View Representation With LSTM for 3-D Shape Recognition and Retrieval

Random-Frequency SAR Imaging Based on Compressed Sensing

Infrared Dim and Small Target Detection via Multiple Subspace Learning and Spatial-Temporal Patch-Tensor Model