Scale variation is one of the key challenges in object detection. In this work, we first present a controlled experiment to investigate the effect of receptive fields for scale variation in object detection. Based on the findings from the exploration experiments, we propose a novel Trident Network (TridentNet) aiming to generate scale-specific feature maps with a uniform representational power. We construct a parallel multi-branch architecture in which each branch shares the same transformation parameters but with different receptive fields. Then, we adopt a scale-aware training scheme to specialize each branch by sampling object instances of proper scales for training. As a bonus, a fast approximation version of TridentNet could achieve significant improvements without any additional parameters and computational cost compared with the vanilla detector. On the COCO dataset, our TridentNet with ResNet-101 backbone achieves state-of-the-art single-model results of 48.4 mAP.Codes are available at https://git.io/fj5vR.

Video objection detection (VID) has been a rising research direction in recent years. A central issue of VID is the appearance degradation of video frames caused by fast motion. This problem is essentially ill-posed for a single frame. Therefore, aggregating features from other frames becomes a natural choice. Existing methods rely heavily on optical flow or recurrent neural networks for feature aggregation. However, these methods emphasize more on the temporally nearby frames. In this work, we argue that aggregating features in the full-sequence level will lead to more discriminative and robust features for video object detection. To achieve this goal, we devise a novel Sequence Level Semantics Aggregation (SELSA) module. We further demonstrate the close relationship between the proposed method and the classic spectral clustering method, providing a novel view for understanding the VID problem. We test the proposed method on the ImageNet VID and the EPIC KITCHENS dataset and achieve new state-of-theart results. Our method does not need complicated postprocessing methods such as Seq-NMS or Tubelet rescoring, which keeps the pipeline simple and clean.

Microtubule-associated protein 1B (MAP1B) is essential for neural development. Besides the abundant expression in neurons, MAP1B recently was found in myelinating oligodendroglia. Moreover, MAP1B deficiency causes delayed myelin development, suggesting the functional importance of MAP1B in oligodendroglia. However, molecular mechanisms that control MAP1B expression in oligodendroglia remain elusive. We report here that MAP1B mRNA is markedly upregulated in the oligodendroglia cell line CG4 upon induced differentiation, leading to elevated MAP1B protein production. A coordinated regulation of homeoprotein transcription factors was observed during CG4 cell differentiation, which recapitulates the regulation in neurons that promotes MAP1B transcription. Hence, transcriptional regulation of MAP1B appears to be a common mechanism in both neurons and oligodendroglia. In addition, we found posttranscriptional regulation of MAP1B mRNA by the selective RNA-binding protein QKI in oligodendroglia. The 3UTR of MAP1B mRNA interacts with QKI, and oligodendroglia-specific QKI-deficiency in the quakingviable mutant mice resulted in reduced MAP1B mRNA expression. Moreover, RNAi-mediated QKI-knockdown caused destabilization of the MAP1B mRNA in CG4 cells. Furthermore, forced expression of exogenous QKI was sufficient for promoting MAP1B expression. Because QKI is absent in neurons, QKI-dependent stabilization of MAP1B mRNA provides a novel mechanism for advancing MAP1B expression specifically in oligodendroglia during brain development. INTRODUCTIONMicrotubule-associated proteins (MAPs) control the dynamic organization of microtubule cytoskeleton, which in turn governs normal cell growth and development (Takemura et al., 1992;Hirokawa, 1994). Among these MAPs, MAP1B is predominantly expressed in the nervous system and is the earliest MAP detected in the embryonic brain (Tucker et al., 1989;Ma et al., 1997;Ohyu et al., 1997). Historically, MAP1B, which has been studied mainly in the developing neurons, plays essential roles in neurite outgrowth, axonal extension, and path finding (Gonzalez-Billault et al., 2001, 2002Bouquet et al., 2004). MAP1B expression is markedly upregulated during neurite outgrowth in various types of neurons (Gordon-Weeks and Fischer, 2000), and MAP1B knockout mice exhibit a range of abnormalities in axonal extension and path finding (Meixner et al., 2000;Gonzalez-Billault et al., 2001;Bouquet et al., 2004). More recent studies indicated that MAP1B expression is not restricted in neurons, but also is detected in oligodendrocytes and Schwann cells that produce myelin in the central and peripheral nervous system (CNS and PNS), respectively (Fischer et al., 1990;Ma et al., 1999). In particular, MAP1B expression is elevated in oligodendrocytes that initiate ensheathment of neuronal axons during normal brain development (Wu et al., 2001) as well as in Schwann cells during nerve regeneration (Ma et al., 1999).The functional importance of MAP1B in CNS myelination is further reinforced by the defects of myelin devel...

No abstract

With the surge of deep learning techniques, the field of person re-identification has witnessed rapid progress in recent years. Deep learning based methods focus on learning a feature space where samples are clustered compactly according to their corresponding identities. Most existing methods rely on powerful CNNs to transform the samples individually. In contrast, we propose to consider the sample relations in the transformation. To achieve this goal, we incorporate spectral clustering technique into CNN. We derive a novel module named Spectral Feature Transformation and seamlessly integrate it into existing CNN pipeline with negligible cost, which makes our method enjoy the best of two worlds. Empirical studies show that the proposed approach outperforms previous state-of-the-art methods on four public benchmarks by a considerable margin without bells and whistles.