Self-supervised learning (SSL) is a popular research topic in speech processing. Successful SSL speech models must generalize well. SUPERB was proposed to evaluate the ability of SSL speech models across many speech tasks. However, due to the diversity of tasks, the evaluation process requires huge computational costs. We present MiniSUPERB, a lightweight benchmark that efficiently evaluates SSL speech models with comparable results to SUPERB while greatly reducing the computational cost. We select representative tasks and sample datasets and extract model representation offline, achieving 0.954 and 0.982 Spearman's rank correlation with SUPERB Paper and SUPERB Challenge, respectively. In the meanwhile, the computational cost is reduced by 97% in regard to MACs (number of Multiply-ACcumulate operations) in the tasks we choose. To the best of our knowledge, this is the first study to examine not only the computational cost of a model itself but the cost of evaluating it on a benchmark.
To achieve accurate 3D object detection at a low cost for autonomous driving, many multi-camera methods have been proposed and solved the occlusion problem of monocular approaches. However, due to the lack of accurate estimated depth, existing multi-camera methods often generate multiple bounding boxes along a ray of depth direction for difficult small objects such as pedestrians, resulting in an extremely low recall. Furthermore, directly applying depth prediction modules to existing multi-camera methods, generally composed of large network architectures, cannot meet the real-time requirements of self-driving applications. To address these issues, we propose Cross-view and Depth-guided Transformers for 3D Object Detection, CrossDTR. First, our lightweight depth predictor is designed to produce precise object-wise sparse depth maps and low-dimensional depth embeddings without extra depth datasets during supervision. Second, a cross-view depth-guided transformer is developed to fuse the depth embeddings as well as image features from cameras of different views and generate 3D bounding boxes. Extensive experiments demonstrated that our method hugely surpassed existing multi-camera methods by 10 percent in pedestrian detection and about 3 percent in overall mAP and NDS metrics. Also, computational analyses showed that our method is 5 times faster than prior approaches. Our codes will be made publicly available at https://github.com/sty61010/CrossDTR.
Lung cancer is one of the deadliest cancers, and in part its effective diagnosis and treatment depend on the accurate delineation of the tumor. Human-centered segmentation, which is currently the most common approach, is subject to inter-observer variability, and is also time-consuming, considering the fact that only experts are capable of providing annotations. Automatic and semi-automatic tumor segmentation methods have recently shown promising results. However, as different researchers have validated their algorithms using various datasets and performance metrics, reliably evaluating these methods is still an open challenge. The goal of the Lung-Originated Tumor Segmentation from Computed Tomography Scan (LOTUS) Benchmark created through 2018 IEEE Video and Image Processing (VIP) Cup competition, is to provide a unique dataset and pre-defined metrics, so that different researchers can develop and evaluate their methods in a unified fashion. The 2018 VIP Cup started with a global engagement from 42 countries to access the competition data. At the registration stage, there were 129 members clustered into 28 teams from 10 countries, out of which 9 teams made it to the final stage and 6 teams successfully completed all the required tasks. In a nutshell, all the algorithms proposed during the competition, are based on deep learning models combined with a false positive reduction technique. Methods developed by the three finalists show promising results in tumor segmentation, however, more effort should be put into reducing the false positive rate. This competition manuscript presents an overview of the VIP-Cup challenge, along with the proposed algorithms and results.
Spatial-temporal prediction is a critical problem for intelligent transportation, which is helpful for tasks such as traffic control and accident prevention. Previous studies rely on large-scale traffic data collected from sensors. However, it is unlikely to deploy sensors in all regions due to the device and maintenance costs. This paper addresses the problem via outdoor cellular traffic distilled from over two billion records per day in a telecom company, because outdoor cellular traffic induced by user mobility is highly related to transportation traffic. We study road intersections in urban and aim to predict future outdoor cellular traffic of all intersections given historic outdoor cellular traffic. Furthermore, We propose a new model for multivariate spatial-temporal prediction, mainly consisting of two extending graph attention networks (GAT). First GAT is used to explore correlations among multivariate cellular traffic. Another GAT leverages the attention mechanism into graph propagation to increase the efficiency of capturing spatial dependency. Experiments show that the proposed model significantly outperforms the state-of-the-art methods on our dataset.
We proposed an end-to-end grasp detection network, Grasp Detection Network (GDN), cooperated with a novel coarse-to-fine (C2F) grasp representation design to detect diverse and accurate 6-DoF grasps based on point clouds. Compared to previous two-stage approaches which sample and evaluate multiple grasp candidates, our architecture is at least 20 times faster. It is also 8% and 40% more accurate in terms of the success rate in single object scenes and the complete rate in clutter scenes, respectively. Our method shows superior results among settings with different number of views and input points. Moreover, we propose a new AP-based metric which considers both rotation and transition errors, making it a more comprehensive evaluation tool for grasp detection models.
With the development of image segmentation in computer vision, biomedical image segmentation have achieved remarkable progress on brain tumor segmentation and Organ At Risk (OAR) segmentation. However, most of the research only uses single modality such as Computed Tomography (CT) scans while in real world scenario doctors often use multiple modalities to get more accurate result. To better leverage different modalities, we have collected a large dataset consists of 136 cases with CT and MR images which diagnosed with nasopharyngeal cancer. In this paper, we propose to use Generative Adversarial Network to perform CT to MR transformation to synthesize MR images instead of aligning two modalities. The synthesized MR can be jointly trained with CT to achieve better performance. In addition, we use instance segmentation model to extend the OAR segmentation task to segment both organs and tumor region. The collected dataset will be made public soon.
How to efficiently utilize temporal information to recover videos in a consistent way is the main issue for video inpainting problems. Conventional 2D CNNs have achieved good performance on image inpainting but often lead to temporally inconsistent results where frames will flicker when applied to videos (see https://www.youtube.com/watch?v=87Vh1HDBjD0&list=PLPoVtv-xp_dL5uckIzz1PKwNjg1yI0I94&index=1); 3D CNNs can capture temporal information but are computationally intensive and hard to train. In this paper, we present a novel component termed Learnable Gated Temporal Shift Module (LGTSM) for video inpainting models that could effectively tackle arbitrary video masks without additional parameters from 3D convolutions. LGTSM is designed to let 2D convolutions make use of neighboring frames more efficiently, which is crucial for video inpainting. Specifically, in each layer, LGTSM learns to shift some channels to its temporal neighbors so that 2D convolutions could be enhanced to handle temporal information. Meanwhile, a gated convolution is applied to the layer to identify the masked areas that are poisoning for conventional convolutions. On the FaceForensics and Free-form Video Inpainting (FVI) dataset, our model achieves state-of-the-art results with simply 33% of parameters and inference time.
Video Question Answering (Video QA) is a critical and challenging task in multimedia comprehension. While deep learning based models are extremely capable of representing and understanding videos, these models heavily rely on massive data, which is expensive to label. In this paper, we introduce a novel task for automatically generating questions given a sequence of video frames and the corresponding subtitles from a clip of video to reduce the huge annotation cost. Learning to ask a question based on a video requires the model to comprehend the rich semantics in the scene and the interplay between the vision and the language. To address this, we propose a novel cross-modal self-attention (CMSA) network to aggregate the diverse features from video frames and subtitles. Excitingly, we demonstrate that our proposed model can improve the (strong) baseline from 0.0738 to 0.1374 in BLEU4 score -- more than 0.063 improvement (i.e., 85\% relatively). Most of all, We arguably pave a novel path toward solving the challenging Video QA task and provide detailed analysis which ushers the avenues for future investigations.