We present an end-to-end deep learning architecture for depth map inference from multi-view images. In the network, we first extract deep visual image features, and then build the 3D cost volume upon the reference camera frustum via the differentiable homography warping. Next, we apply 3D convolutions to regularize and regress the initial depth map, which is then refined with the reference image to generate the final output. Our framework flexibly adapts arbitrary N-view inputs using a variance-based cost metric that maps multiple features into one cost feature. The proposed MVSNet is demonstrated on the large-scale indoor DTU dataset. With simple post-processing, our method not only significantly outperforms previous state-of-the-arts, but also is several times faster in runtime. We also evaluate MVSNet on the complex outdoor Tanks and Temples dataset, where our method ranks first before April 18, 2018 without any fine-tuning, showing the strong generalization ability of MVSNet.

Deep learning has recently demonstrated its excellent performance for multi-view stereo (MVS). However, one major limitation of current learned MVS approaches is the scalability: the memory-consuming cost volume regularization makes the learned MVS hard to be applied to highresolution scenes. In this paper, we introduce a scalable multi-view stereo framework based on the recurrent neural network. Instead of regularizing the entire 3D cost volume in one go, the proposed Recurrent Multi-view Stereo Network (R-MVSNet) sequentially regularizes the 2D cost maps along the depth direction via the gated recurrent unit (GRU). This reduces dramatically the memory consumption and makes high-resolution reconstruction feasible. We first show the state-of-the-art performance achieved by the proposed R-MVSNet on the recent MVS benchmarks. Then, we further demonstrate the scalability of the proposed method on several large-scale scenarios, where previous learned approaches often fail due to the memory constraint. Code is available at https://github.com/ YoYo000/MVSNet.

No abstract

The reconstruction of neuronal populations, a key step in understanding neural circuits, remains a challenge in the presence of densely packed neurites. Here we achieved automatic reconstruction of neuronal populations by partially mimicking human strategies to separate individual neurons. For populations not resolvable by other methods, we obtained recall and precision rates of approximately 80%. We also demonstrate the reconstruction of 960 neurons within 3 h.

No abstract

No abstract

Drawing the map of neuronal circuits at microscopic resolution is important to explain how brain works. Recent progresses in fluorescence labeling and imaging techniques have enabled measuring the whole brain of a rodent like a mouse at submicron-resolution. Considering the huge volume of such datasets, automatic tracing and reconstruct the neuronal connections from the image stacks is essential to form the large scale circuits. However, the first step among which, automated location the soma across different brain areas remains a challenge. Here, we addressed this problem by introducing L1 minimization model. We developed a fully automated system, NeuronGlobalPositionSystem (NeuroGPS) that is robust to the broad diversity of shape, size and density of the neurons in a mouse brain. This method allows locating the neurons across different brain areas without human intervention. We believe this method would facilitate the analysis of the neuronal circuits for brain function and disease studies.

No abstract