Given a single photo of a room and a large database of furniture CAD models, our goal is to reconstruct a scene that is as similar as possible to the scene depicted in the photograph, and composed of objects drawn from the database. We present a completely automatic system to address this IM2CAD problem that produces high quality results on challenging imagery from interior home design and remodeling websites. Our approach iteratively optimizes the placement and scale of objects in the room to best match scene renderings to the input photo, using image comparison metrics trained via deep convolutional neural nets. By operating jointly on the full scene at once, we account for inter-object occlusions. We also show the applicability of our method in standard scene understanding benchmarks where we obtain significant improvement.
Recently, deep convolutional neural network (DCNN) achieved increasingly remarkable success and rapidly developed in the field of natural image recognition. Compared with the natural image, the scale of remote sensing image is larger and the scene and the object it represents are more macroscopic. This study inquires whether remote sensing scene and natural scene recognitions differ and raises the following questions: What are the key factors in remote sensing scene recognition? Is the DCNN recognition mechanism centered on object recognition still applicable to the scenarios of remote sensing scene understanding? We performed several experiments to explore the influence of the DCNN structure and the scale of remote sensing scene understanding from the perspective of scene complexity. Our experiment shows that understanding a complex scene depends on an in-depth network and multiple-scale perception. Using a visualization method, we qualitatively and quantitatively analyze the recognition mechanism in a complex remote sensing scene and demonstrate the importance of multi-objective joint semantic support.
In this paper, we discuss and review how combined multi-view imagery from satellite to street-level can benefit scene analysis. Numerous works exist that merge information from remote sensing and images acquired from the ground for tasks like land cover mapping, object detection, or scene understanding. What makes the combination of overhead and street-level images challenging, is the strongly varying viewpoint, different scale, illumination, sensor modality and time of acquisition. Direct (dense) matching of images on a per-pixel basis is thus often impossible, and one has to resort to alternative strategies that will be discussed in this paper. We review recent works that attempt to combine images taken from the ground and overhead views for purposes like scene registration, reconstruction, or classification. Three methods that represent the wide range of potential methods and applications (change detection, image orientation, and tree cataloging) are described in detail. We show that cross-fertilization between remote sensing, computer vision and machine learning is very valuable to make the best of geographic data available from Earth Observation sensors and ground imagery. Despite its challenges, we believe that integrating these complementary data sources will lead to major breakthroughs in Big GeoData.
We present a powerful method to extract per-point semantic class labels from aerialphotogrammetry data. Labeling this kind of data is important for tasks such as environmental modelling, object classification and scene understanding. Unlike previous point cloud classification methods that rely exclusively on geometric features, we show that incorporating color information yields a significant increase in accuracy in detecting semantic classes. We test our classification method on three real-world photogrammetry datasets that were generated with Pix4Dmapper Pro, and with varying point densities. We show that off-the-shelf machine learning techniques coupled with our new features allow us to train highly accurate classifiers that generalize well to unseen data, processing point clouds containing 10 million points in less than 3 minutes on a desktop computer.
In the context of scene understanding, a variety of methods exists to estimate different information channels from mono or stereo images, including disparity, depth, and normals. Although several advances have been reported in the recent years for these tasks, the estimated information is often imprecise particularly near depth discontinuities or creases. Studies have however shown that precisely such depth edges carry critical cues for the perception of shape, and play important roles in tasks like depth-based segmentation or foreground selection. Unfortunately, the currently extracted channels often carry conflicting signals, making it difficult for subsequent applications to effectively use them. In this paper, we focus on the problem of obtaining high-precision depth edges (i.e., depth contours and creases) by jointly analyzing such unreliable information channels. We propose DepthCut, a data-driven fusion of the channels using a convolutional neural network trained on a large dataset with known depth. The resulting depth edges can be used for segmentation, decomposing a scene into depth layers with relatively flat depth, or improving the accuracy of the depth estimate near depth edges by constraining its gradients to agree with these edges. Quantitatively, we compare against 15 variants of baselines and demonstrate that our depth edges result in an improved segmentation performance and an improved depth estimate near depth edges compared to data-agnostic channel fusion. Qualitatively, we demonstrate that the depth edges result in superior segmentation and depth orderings.
Convolutional networks for image classification progressively reduce resolution until the image is represented by tiny feature maps in which the spatial structure of the scene is no longer discernible. Such loss of spatial acuity can limit image classification accuracy and complicate the transfer of the model to downstream applications that require detailed scene understanding. These problems can be alleviated by dilation, which increases the resolution of output feature maps without reducing the receptive field of individual neurons. We show that dilated residual networks (DRNs) outperform their non-dilated counterparts in image classification without increasing the model's depth or complexity. We then study gridding artifacts introduced by dilation, develop an approach to removing these artifacts (`degridding'), and show that this further increases the performance of DRNs. In addition, we show that the accuracy advantage of DRNs is further magnified in downstream applications such as object localization and semantic segmentation.
3D scene understanding is important for robots to interact with the 3D world in a meaningful way. Most previous works on 3D scene understanding focus on recognizing geometrical or semantic properties of the scene independently. In this work, we introduce Data Associated Recurrent Neural Networks (DA-RNNs), a novel framework for joint 3D scene mapping and semantic labeling. DA-RNNs use a new recurrent neural network architecture for semantic labeling on RGB-D videos. The output of the network is integrated with mapping techniques such as KinectFusion in order to inject semantic information into the reconstructed 3D scene. Experiments conducted on a real world dataset and a synthetic dataset with RGB-D videos demonstrate the ability of our method in semantic 3D scene mapping.
Pixel-wise semantic segmentation for visual scene understanding not only needs to be accurate, but also efficient in order to find any use in real-time application. Existing algorithms even though are accurate but they do not focus on utilizing the parameters of neural network efficiently. As a result they are huge in terms of parameters and number of operations; hence slow too. In this paper, we propose a novel deep neural network architecture which allows it to learn without any significant increase in number of parameters. Our network uses only 11.5 million parameters and 21.2 GFLOPs for processing an image of resolution 3x640x360. It gives state-of-the-art performance on CamVid and comparable results on Cityscapes dataset. We also compare our networks processing time on NVIDIA GPU and embedded system device with existing state-of-the-art architectures for different image resolutions.
Indoor scene understanding is central to applications such as robot navigation and human companion assistance. Over the last years, data-driven deep neural networks have outperformed many traditional approaches thanks to their representation learning capabilities. One of the bottlenecks in training for better representations is the amount of available per-pixel ground truth data that is required for core scene understanding tasks such as semantic segmentation, normal prediction, and object edge detection. To address this problem, a number of works proposed using synthetic data. However, a systematic study of how such synthetic data is generated is missing. In this work, we introduce a large-scale synthetic dataset with 400K physically-based rendered images from 45K realistic 3D indoor scenes. We study the effects of rendering methods and scene lighting on training for three computer vision tasks: surface normal prediction, semantic segmentation, and object boundary detection. This study provides insights into the best practices for training with synthetic data (more realistic rendering is worth it) and shows that pretraining with our new synthetic dataset can improve results beyond the current state of the art on all three tasks.
Since the beginning of early civilizations, social relationships derived from each individual fundamentally form the basis of social structure in our daily life. In the computer vision literature, much progress has been made in scene understanding, such as object detection and scene parsing. Recent research focuses on the relationship between objects based on its functionality and geometrical relations. In this work, we aim to study the problem of social relationship recognition, in still images. We have proposed a dual-glance model for social relationship recognition, where the first glance fixates at the individual pair of interest and the second glance deploys attention mechanism to explore contextual cues. We have also collected a new large scale People in Social Context (PISC) dataset, which comprises of 22,670 images and 76,568 annotated samples from 9 types of social relationship. We provide benchmark results on the PISC dataset, and qualitatively demonstrate the efficacy of the proposed model.
The ability to predict and therefore to anticipate the future is an important attribute of intelligence. It is also of utmost importance in real-time systems, e.g. in robotics or autonomous driving, which depend on visual scene understanding for decision making. While prediction of the raw RGB pixel values in future video frames has been studied in previous work, here we introduce the novel task of predicting semantic segmentations of future frames. Given a sequence of video frames, our goal is to predict segmentation maps of not yet observed video frames that lie up to a second or further in the future. We develop an autoregressive convolutional neural network that learns to iteratively generate multiple frames. Our results on the Cityscapes dataset show that directly predicting future segmentations is substantially better than predicting and then segmenting future RGB frames. Prediction results up to half a second in the future are visually convincing and are much more accurate than those of a baseline based on warping semantic segmentations using optical flow.
Real-time scene understanding has become crucial in many applications such as autonomous driving. In this paper, we propose a deep architecture, called BlitzNet, that jointly performs object detection and semantic segmentation in one forward pass, allowing real-time computations. Besides the computational gain of having a single network to perform several tasks, we show that object detection and semantic segmentation benefit from each other in terms of accuracy. Experimental results for VOC and COCO datasets show state-of-the-art performance for object detection and segmentation among real time systems.
Most approaches for instance-aware semantic labeling traditionally focus on accuracy. Other aspects like runtime and memory footprint are arguably as important for real-time applications such as autonomous driving. Motivated by this observation and inspired by recent works that tackle multiple tasks with a single integrated architecture, in this paper we present a real-time efficient implementation based on ENet that solves three autonomous driving related tasks at once: semantic scene segmentation, instance segmentation and monocular depth estimation. Our approach builds upon a branched ENet architecture with a shared encoder but different decoder branches for each of the three tasks. The presented method can run at 21 fps at a resolution of 1024x512 on the Cityscapes dataset without sacrificing accuracy compared to running each task separately.
Depth estimation from single monocular images is a key component of scene understanding and has benefited largely from deep convolutional neural networks (CNN) recently. In this article, we take advantage of the recent deep residual networks and propose a simple yet effective approach to this problem. We formulate depth estimation as a pixel-wise classification task. Specifically, we first discretize the continuous depth values into multiple bins and label the bins according to their depth range. Then we train fully convolutional deep residual networks to predict the depth label of each pixel. Performing discrete depth label classification instead of continuous depth value regression allows us to predict a confidence in the form of probability distribution. We further apply fully-connected conditional random fields (CRF) as a post processing step to enforce local smoothness interactions, which improves the results. We evaluate our approach on both indoor and outdoor datasets and achieve state-of-the-art performance.
A growing demand for natural-scene text detection has been witnessed by the computer vision community since text information plays a significant role in scene understanding and image indexing. Deep neural networks are being used due to their strong capabilities of pixel-wise classification or word localization, similar to being used in common vision problems. In this paper, we present a novel two-task network with integrating bottom and top cues. The first task aims to predict a pixel-by-pixel labeling and based on which, word proposals are generated with a canonical connected component analysis. The second task aims to output a bundle of character candidates used later to verify the word proposals. The two sub-networks share base convolutional features and moreover, we present a new loss to strengthen the interaction between them. We evaluate the proposed network on public benchmark datasets and show it can detect arbitrary-orientation scene text with a finer output boundary. In ICDAR 2013 text localization task, we achieve the state-of-the-art performance with an F-score of 0.919 and a much better recall of 0.915.
While deep neural networks have led to human-level performance on computer vision tasks, they have yet to demonstrate similar gains for holistic scene understanding. In particular, 3D context has been shown to be an extremely important cue for scene understanding - yet very little research has been done on integrating context information with deep models. This paper presents an approach to embed 3D context into the topology of a neural network trained to perform holistic scene understanding. Given a depth image depicting a 3D scene, our network aligns the observed scene with a predefined 3D scene template, and then reasons about the existence and location of each object within the scene template. In doing so, our model recognizes multiple objects in a single forward pass of a 3D convolutional neural network, capturing both global scene and local object information simultaneously. To create training data for this 3D network, we generate partly hallucinated depth images which are rendered by replacing real objects with a repository of CAD models of the same object category. Extensive experiments demonstrate the effectiveness of our algorithm compared to the state-of-the-arts. Source code and data are available at http://deepcontext.cs.princeton.edu.
3D pose estimation is a key component of many important computer vision tasks such as autonomous navigation and 3D scene understanding. Most state-of-the-art approaches to 3D pose estimation solve this problem as a pose-classification problem in which the pose space is discretized into bins and a CNN classifier is used to predict a pose bin. We argue that the 3D pose space is continuous and propose to solve the pose estimation problem in a CNN regression framework with a suitable representation, data augmentation and loss function that captures the geometry of the pose space. Experiments on PASCAL3D+ show that the proposed 3D pose regression approach achieves competitive performance compared to the state-of-the-art.
Visual information plays an indispensable role in our daily interactions with environment. Such information is manipulated for a wide range of purposes spanning from basic object and material perception to complex gesture interpretations. There have been novel studies in cognitive science for in-depth understanding of visual information manipulation, which lead to answer questions such as: how we infer 2D/3D motion from a sequence of 2D images? how we understand a motion from a single image frame? how we see forest avoiding trees? Leveraging on congruence, linear mapping transformation determination between a set of shapes facilitate motion perception. Present study methodizes recent discoveries of human cognitive ability for scene understanding. The proposed method processes images hierarchically, that is an iterative analysis of scene abstractions using a rapidly converging heuristic iterative method. The method hierarchically abstracts images; the abstractions are represented in polar coordinate system, and any two consecutive abstractions have incremental level of details. The method then creates a graph of approximated linear mapping transformations based on circular shift permutations of hierarchical abstractions. The graph is then traversed in best-first fashion to find best linear mapping transformation. The accuracy of the proposed method is assessed using normal, noisy, and deformed images. Additionally, the present study deduces (i) the possibility of determining optimal mapping linear transformations in logarithmic iterations with respect to the precision of results, and (ii) computational cost is independent from the resolution of input shapes.
This work addresses the problem of semantic foggy scene understanding (SFSU). Although extensive research has been performed on image dehazing and on semantic scene understanding with clear-weather images, little attention has been paid to SFSU. Due to the difficulty of collecting and annotating foggy images, we choose to generate synthetic fog on real images that depict clear-weather outdoor scenes, and then leverage these partially synthetic data for SFSU by employing state-of-the-art convolutional neural networks (CNN). In particular, a complete pipeline to add synthetic fog to real, clear-weather images using incomplete depth information is developed. We apply our fog synthesis on the Cityscapes dataset and generate Foggy Cityscapes with 20550 images. SFSU is tackled in two ways: 1) with typical supervised learning, and 2) with a novel type of semi-supervised learning, which combines 1) with an unsupervised supervision transfer from clear-weather images to their synthetic foggy counterparts. In addition, we carefully study the usefulness of image dehazing for SFSU. For evaluation, we present Foggy Driving, a dataset with 101 real-world images depicting foggy driving scenes, which come with ground truth annotations for semantic segmentation and object detection. Extensive experiments show that 1) supervised learning with our synthetic data significantly improves the performance of state-of-the-art CNN for SFSU on Foggy Driving; 2) our semi-supervised learning strategy further improves performance; and 3) image dehazing marginally advances SFSU with our learning strategy. The datasets, models and code are made publicly available.
We present a novel data set made up of omnidirectional video of multiple objects whose centroid positions are annotated automatically. Omnidirectional vision is an active field of research focused on the use of spherical imagery in video analysis and scene understanding, involving tasks such as object detection, tracking and recognition. Our goal is to provide a large and consistently annotated video data set that can be used to train and evaluate new algorithms for these tasks. Here we describe the experimental setup and software environment used to capture and map the 3D ground truth positions of multiple objects into the image. Furthermore, we estimate the expected systematic error on the mapped positions. In addition to final data products, we release publicly the software tools and raw data necessary to re-calibrate the camera and/or redo this mapping. The software also provides a simple framework for comparing the results of standard image annotation tools or visual tracking systems against our mapped ground truth annotations.