In this paper, we propose a generative model, Temporal Generative Adversarial Nets (TGAN), which can learn a semantic representation of unlabeled videos, and is capable of generating videos. Unlike existing Generative Adversarial Nets (GAN)-based methods that generate videos with a single generator consisting of 3D deconvolutional layers, our model exploits two different types of generators: a temporal generator and an image generator. The temporal generator takes a single latent variable as input and outputs a set of latent variables, each of which corresponds to an image frame in a video. The image generator transforms a set of such latent variables into a video. To deal with instability in training of GAN with such advanced networks, we adopt a recently proposed model, Wasserstein GAN, and propose a novel method to train it stably in an end-to-end manner. The experimental results demonstrate the effectiveness of our methods.

No abstract

No abstract

No abstract

Building and road detection from aerial imagery has many applications in a wide range of areas including urban design, real-estate management, and disaster relief. The extracting buildings and roads from aerial imagery has been performed by human experts manually, so that it has been very costly and time-consuming process. Our goal is to develop a system for automatically detecting buildings and roads directly from aerial imagery. Many attempts at automatic aerial imagery interpretation have been proposed in remote sensing literature, but much of early works use local features to classify each pixel or segment to an object label, so that these kind of approach needs some prior knowledge on object appearance or class-conditional distribution of pixel values. Furthermore, some works also need a segmentation step as pre-processing. Therefore, we use Convolutional Neural Networks(CNN) to learn mapping from raw pixel values in aerial imagery to three object labels (buildings, roads, and others), in other words, we generate three-channel maps from raw aerial imagery input. We take a patch-based semantic segmentation approach, so we firstly divide large aerial imagery into small patches and then train the CNN with those patches and corresponding three-channel map patches. Finally, we evaluate our system on a large-scale road and building detection datasets that is publicly available.

We present an approach for road segmentation that only requires image-level annotations at training time. We leverage distant supervision, which allows us to train our model using images that are different from the target domain. Using large publicly available image databases as distant supervisors, we develop a simple method to automatically generate weak pixel-wise road masks. These are used to iteratively train a fully convolutional neural network, which produces our final segmentation model. We evaluate our method on the Cityscapes dataset, where we compare it with a fully supervised approach. Further, we discuss the tradeoff between annotation cost and performance. Overall, our distantly supervised approach achieves 93.8% of the performance of the fully supervised approach, while using orders of magnitude less annotation work.

No abstract

Identifying "free-space," or safely driveable regions in the scene ahead, is a fundamental task for autonomous navigation. While this task can be addressed using semantic segmentation, the manual labor involved in creating pixelwise annotations to train the segmentation model is very costly. Although weakly supervised segmentation addresses this issue, most methods are not designed for free-space. In this paper, we observe that homogeneous texture and location are two key characteristics of free-space, and develop a novel, practical framework for free-space segmentation with minimal human supervision. Our experiments show that our framework performs better than other weakly supervised methods while using less supervision. Our work demonstrates the potential for performing free-space segmentation without tedious and costly manual annotation, which will be important for adapting autonomous driving systems to different types of vehicles and environments.