Video object detection is challenging in the presence of appearance deterioration in certain video frames. Therefore, it is a natural choice to aggregate temporal information from other frames of the same video into the current frame. However, RoI Al
ign, as one of the most core procedures of video detectors, still remains extracting features from a single-frame feature map for proposals, making the extracted RoI features lack temporal information from videos. In this work, considering the features of the same object instance are highly similar among frames in a video, a novel Temporal RoI Align operator is proposed to extract features from other frames feature maps for current frame proposals by utilizing feature similarity. The proposed Temporal RoI Align operator can extract temporal information from the entire video for proposals. We integrate it into single-frame video detectors and other state-of-the-art video detectors, and conduct quantitative experiments to demonstrate that the proposed Temporal RoI Align operator can consistently and significantly boost the performance. Besides, the proposed Temporal RoI Align can also be applied into video instance segmentation. Codes are available at https://github.com/open-mmlab/mmtracking
The performance of object detection, to a great extent, depends on the availability of large annotated datasets. To alleviate the annotation cost, the research community has explored a number of ways to exploit unlabeled or weakly labeled data. Howev
er, such efforts have met with limited success so far. In this work, we revisit the problem with a pragmatic standpoint, trying to explore a new balance between detection performance and annotation cost by jointly exploiting fully and weakly annotated data. Specifically, we propose a weakly- and semi-supervised object detection framework (WSSOD), which involves a two-stage learning procedure. An agent detector is first trained on a joint dataset and then used to predict pseudo bounding boxes on weakly-annotated images. The underlying assumptions in the current as well as common semi-supervised pipelines are also carefully examined under a unified EM formulation. On top of this framework, weakly-supervised loss (WSL), label attention and random pseudo-label sampling (RPS) strategies are introduced to relax these assumptions, bringing additional improvement on the efficacy of the detection pipeline. The proposed framework demonstrates remarkable performance on PASCAL-VOC and MSCOCO benchmark, achieving a high performance comparable to those obtained in fully-supervised settings, with only one third of the annotations.
Controversy exists on whether differentiable neural architecture search methods discover wiring topology effectively. To understand how wiring topology evolves, we study the underlying mechanism of several existing differentiable NAS frameworks. Our
investigation is motivated by three observed searching patterns of differentiable NAS: 1) they search by growing instead of pruning; 2) wider networks are more preferred than deeper ones; 3) no edges are selected in bi-level optimization. To anatomize these phenomena, we propose a unified view on searching algorithms of existing frameworks, transferring the global optimization to local cost minimization. Based on this reformulation, we conduct empirical and theoretical analyses, revealing implicit inductive biases in the costs assignment mechanism and evolution dynamics that cause the observed phenomena. These biases indicate strong discrimination towards certain topologies. To this end, we pose questions that future differentiable methods for neural wiring discovery need to confront, hoping to evoke a discussion and rethinking on how much bias has been enforced implicitly in existing NAS methods.
Feature pyramid has been an efficient method to extract features at different scales. Development over this method mainly focuses on aggregating contextual information at different levels while seldom touching the inter-level correlation in the featu
re pyramid. Early computer vision methods extracted scale-invariant features by locating the feature extrema in both spatial and scale dimension. Inspired by this, a convolution across the pyramid level is proposed in this study, which is termed pyramid convolution and is a modified 3-D convolution. Stacked pyramid convolutions directly extract 3-D (scale and spatial) features and outperforms other meticulously designed feature fusion modules. Based on the viewpoint of 3-D convolution, an integrated batch normalization that collects statistics from the whole feature pyramid is naturally inserted after the pyramid convolution. Furthermore, we also show that the naive pyramid convolution, together with the design of RetinaNet head, actually best applies for extracting features from a Gaussian pyramid, whose properties can hardly be satisfied by a feature pyramid. In order to alleviate this discrepancy, we build a scale-equalizing pyramid convolution (SEPC) that aligns the shared pyramid convolution kernel only at high-level feature maps. Being computationally efficient and compatible with the head design of most single-stage object detectors, the SEPC module brings significant performance improvement ($>4$AP increase on MS-COCO2017 dataset) in state-of-the-art one-stage object detectors, and a light version of SEPC also has $sim3.5$AP gain with only around 7% inference time increase. The pyramid convolution also functions well as a stand-alone module in two-stage object detectors and is able to improve the performance by $sim2$AP. The source code can be found at https://github.com/jshilong/SEPC.
Channel pruning is a popular technique for compressing convolutional neural networks (CNNs), where various pruning criteria have been proposed to remove the redundant filters. From our comprehensive experiments, we found two blind spots in the study
of pruning criteria: (1) Similarity: There are some strong similarities among several primary pruning criteria that are widely cited and compared. According to these criteria, the ranks of filtersImportance Score are almost identical, resulting in similar pruned structures. (2) Applicability: The filtersImportance Score measured by some pruning criteria are too close to distinguish the network redundancy well. In this paper, we analyze these two blind spots on different types of pruning criteria with layer-wise pruning or global pruning. The analyses are based on the empirical experiments and our assumption (Convolutional Weight Distribution Assumption) that the well-trained convolutional filters each layer approximately follow a Gaussian-alike distribution. This assumption has been verified through systematic and extensive statistical tests.
Improving sparsity of deep neural networks (DNNs) is essential for network compression and has drawn much attention. In this work, we disclose a harmful sparsifying process called filter collapse, which is common in DNNs with batch normalization (BN)
and rectified linear activation functions (e.g. ReLU, Leaky ReLU). It occurs even without explicit sparsity-inducing regularizations such as $L_1$. This phenomenon is caused by the normalization effect of BN, which induces a non-trainable region in the parameter space and reduces the network capacity as a result. This phenomenon becomes more prominent when the network is trained with large learning rates (LR) or adaptive LR schedulers, and when the network is finetuned. We analytically prove that the parameters of BN tend to become sparser during SGD updates with high gradient noise and that the sparsifying probability is proportional to the square of learning rate and inversely proportional to the square of the scale parameter of BN. To prevent the undesirable collapsed filters, we propose a simple yet effective approach named post-shifted BN (psBN), which has the same representation ability as BN while being able to automatically make BN parameters trainable again as they saturate during training. With psBN, we can recover collapsed filters and increase the model performance in various tasks such as classification on CIFAR-10 and object detection on MS-COCO2017.
In the last decade the ns$^2$ cations (e.g., Pb$^{2+}$ and Sn$^{2+}$) based halides have emerged as one of the most exciting new classes of optoelectronic materials, as exemplified by for instance hybrid perovskite solar absorbers. These materials no
t only exhibit unprecedented performance in some cases, but they also appear to break new ground with their unexpected properties, such as extreme tolerance to defects. However, because of the relatively recent emergence of this class of materials, there remain many yet to be fully explored compounds. Here we assess a series of bismuth/antimony oxyhalides and chalcohalides using consistent first principles methods to ascertain their properties and obtain trends. Based on these calculations, we identify a subset consisting of three types of compounds that may be promising as solar absorbers, transparent conductors, and radiation detectors. Their electronic structure, connection to the crystal geometry, and impact on band-edge dispersion and carrier effective mass are discussed.
