We present MMDetection, an object detection toolbox that contains a rich set of object detection and instance segmentation methods as well as related components and modules. The toolbox started from a codebase of MMDet team who won the detection track of COCO Challenge 2018. It gradually evolves into a unified platform that covers many popular detection methods and contemporary modules. It not only includes training and inference codes, but also provides weights for more than 200 network models. We believe this toolbox is by far the most complete detection toolbox. In this paper, we introduce the various features of this toolbox. In addition, we also conduct a benchmarking study on different methods, components, and their hyper-parameters. We wish that the toolbox and benchmark could serve the growing research community by providing a flexible toolkit to reimplement existing methods and develop their own new detectors. Code and models are available at https: //github.com/open-mmlab/mmdetection. The project is under active development and we will keep this document updated.

Cascade is a classic yet powerful architecture that has boosted performance on various tasks. However, how to introduce cascade to instance segmentation remains an open question. A simple combination of Cascade R-CNN and Mask R-CNN only brings limited gain. In exploring a more effective approach, we find that the key to a successful instance segmentation cascade is to fully leverage the reciprocal relationship between detection and segmentation. In this work, we propose a new framework, Hybrid Task Cascade (HTC), which differs in two important aspects: (1) instead of performing cascaded refinement on these two tasks separately, it interweaves them for a joint multi-stage processing; (2) it adopts a fully convolutional branch to provide spatial context, which can help distinguishing hard foreground from cluttered background. Overall, this framework can learn more discriminative features progressively while integrating complementary features together in each stage. Without bells and whistles, a single HTC obtains 38.4% and 1.5% improvement over a strong Cascade Mask R-CNN baseline on MSCOCO dataset. Moreover, our overall system achieves 48.6 mask AP on the test-challenge split, ranking 1st in the COCO 2018 Challenge Object Detection Task. Code is available at: https://github.com/ open-mmlab/mmdetection.

Chemoreceptor arrays are supramolecular transmembrane machines of unknown structure that allow bacteria to sense their surroundings and respond by chemotaxis. We have combined X-ray crystallography of purified proteins with electron cryotomography of native arrays inside cells to reveal the arrangement of the component transmembrane receptors, histidine kinases (CheA) and CheW coupling proteins. Trimers of receptor dimers lie at the vertices of a hexagonal lattice in a "two-facing-two" configuration surrounding a ring of alternating CheA regulatory domains (P5) and CheW couplers. Whereas the CheA kinase domains (P4) project downward below the ring, the CheA dimerization domains (P3) link neighboring rings to form an extended, stable array. This highly interconnected protein architecture underlies the remarkable sensitivity and cooperative nature of transmembrane signaling in bacterial chemotaxis.protein structure | hybrid methods | two-component systems C hemotactic bacteria sense their surrounding conditions through an array of transmembrane chemoreceptors (methylaccepting chemotaxis proteins, or MCPs), which are found with histidine kinases (CheA) and couplers (CheW) in polar clusters (1-3) and along the sides of cells (4, 5). Repellents and attractants bind to the periplasmic domains of the MCPs either directly (6, 7) or via periplasmic binding proteins (8). The status of the binding domain is transmitted along the length of the receptors through the transmembrane region, across one or more HAMP (histidine kinases, adenyl cyclases, MCPs, and some phosphatases) domain (s), and down the coiled-coil cytoplasmic signaling domain where they ultimately regulate the activity of the histidine kinase CheA located at the receptors' cytoplasmic tips (1-3, 9). CheA is a large, five-domain (P1-P5) protein. P1 contains the substrate histidine, P2 is the docking site for the response regulator CheY, P3 is the dimerization domain, P4 binds ATP and is the kinase, and P5 binds CheW. P1, P2, and P3 are connected to each other by flexible linkers (1, 2). Crystal structures of all domains from Thermotoga maritima CheA are already available (10-13).In the model system Escherichia coli, the addition of attractants or removal of repellents results in kinase inactivation, causing the flagella to rotate counterclockwise. In that case, the multiple flagella form one large bundle that propels the cells smoothly forward and the cells "run." In contrast, addition of repellents or removal of attractants activates CheA, which autophosphorylates and then transfers the phosphoryl group to the second messenger CheY, which in turn binds to the flagellar motors and changes the direction of flagellar rotation to clockwise (CW). This switch results in disassembly of the flagellar bundle and causes the cells to "tumble" (14). CheA also regulates the activity of the receptor-modifying enzyme CheB (a methylesterase), which together with CheR (a methyltransferase) controls the methylation state of residues in the MCP adaptation region (1). Methylation...

Electrochemical reduction of carbon dioxide with renewable energy is a sustainable way of producing carbon-neutral fuels. However, developing active, selective and stable electrocatalysts is challenging and entails material structure design and tailoring across a range of length scales. Here we report a cobalt-phthalocyanine-based high-performance carbon dioxide reduction electrocatalyst material developed with a combined nanoscale and molecular approach. On the nanoscale, cobalt phthalocyanine (CoPc) molecules are uniformly anchored on carbon nanotubes to afford substantially increased current density, improved selectivity for carbon monoxide, and enhanced durability. On the molecular level, the catalytic performance is further enhanced by introducing cyano groups to the CoPc molecule. The resulting hybrid catalyst exhibits >95% Faradaic efficiency for carbon monoxide production in a wide potential range and extraordinary catalytic activity with a current density of 15.0 mA cm−2 and a turnover frequency of 4.1 s−1 at the overpotential of 0.52 V in a near-neutral aqueous solution.

This paper addresses semantic image segmentation by incorporating rich information into Markov Random Field (MRF), including high-order relations and mixture of label contexts. Unlike previous works that optimized MRFs using iterative algorithm, we solve MRF by proposing a Convolutional Neural Network (CNN), namely Deep Parsing Network (DPN), which enables deterministic end-toend computation in a single forward pass. Specifically, DPN extends a contemporary CNN architecture to model unary terms and additional layers are carefully devised to approximate the mean field algorithm (MF) for pairwise terms. It has several appealing properties. First, different from the recent works that combined CNN and MRF, where many iterations of MF were required for each training image during back-propagation, DPN is able to achieve high performance by approximating one iteration of MF. Second, DPN represents various types of pairwise terms, making many existing works as its special cases. Third, DPN makes MF easier to be parallelized and speeded up in Graphical Processing Unit (GPU). DPN is thoroughly evaluated on the PASCAL VOC 2012 dataset, where a single DPN model yields a new state-of-the-art segmentation accuracy of 77.5%.