Rich scene understanding from 3-D point clouds is a challenging task that requires contextual reasoning, which is typically computationally expensive. The task is further complicated when we expect the scene analysis algorithm to also efficiently handle data that is continuously streamed from a sensor on a mobile robot. Hence, we are typically forced to make a choice between 1) using a precise representation of the scene at the cost of speed, or 2) making fast, though inaccurate, approximations at the cost of increased misclassifications. In this work, we demonstrate that we can achieve the best of both worlds by using an efficient and simple representation of the scene in conjunction with recent developments in structured prediction in order to obtain both efficient and state-of-the-art classifications. Furthermore, this efficient scene representation naturally handles streaming data and provides a 300% to 500% speedup over more precise representations.

We address the problem of image-based scene analysis from streaming video, as would be seen from a moving platform, in order to efficiently generate spatially and temporally consistent predictions of semantic categories over time. In contrast to previous techniques which typically address this problem in batch and/or through graphical models, we demonstrate that by learning visual similarities between pixels across frames, a simple filtering algorithfiltering algorithmm is able to achieve high performance predictions in an efficient and online/causal manner. Our technique is a meta-algorithm that can be efficiently wrapped around any scene analysis technique that produces a per-pixel semantic category distribution. We validate our approach over three different scene analysis techniques on three different datasets that contain different semantic object categories. Our experiments demonstrate that our approach is very efficient in practice and substantially improves the consistency of the predictions over time.

Nearly every structured prediction problem in computer vision requires approximate inference due to large and complex dependencies among output labels. While graphical models provide a clean separation between modeling and inference, learning these models with approximate inference is not well understood. Furthermore, even if a good model is learned, predictions are often inaccurate due to approximations. In this work, instead of performing inference over a graphical model, we instead consider the inference procedure as a composition of predictors. Specifically, we focus on message-passing algorithms, such as Belief Propagation, and show how they can be viewed as procedures that sequentially predict label distributions at each node over a graph. Given labeled graphs, we can then train the sequence of predictors to output the correct labeling s. The result no longer corresponds to a graphical model but simply defines an inference procedure, with strong theoretical properties, that can be used to classify new graphs. We demonstrate the scalability and efficacy of our approach on 3D point cloud classification and 3D surface estimation from single images.

We address the problem of understanding scenes from 3-D laser scans via per-point assignment of semantic labels. In order to mitigate the difficulties of using a graphical model for modeling the contextual relationships among the 3-D points, we instead propose a multi-stage inference procedure to capture these relationships. More specifically, we train this procedure to use point cloud statistics and learn relational information (e.g., tree-trunks are below vegetation) over fine (point-wise) and coarse (region-wise) scales. We evaluate our approach on three different datasets, that were obtained from different sensors, and demonstrate improved performance.

Multiple input multiple output (MIMO) orthogonal frequency division multiplexing (OFDM) systems facilitate high data rate wireless communications, and require reliable channel estimates to fully materialize their advantages. The semiconductor downscaling trend has exacerbated device impairments such as inphase and quadrature (I/Q) imbalances which cause inter-carrier interferences in OFDM systems which cannot be remedied by increasing signal power. Different RF chains of MIMO branches can cause different I/Q imbalances which further complicates MIMO OFDM channel estimation. This paper proposes several pilot designs for the estimation of the combined responses of MIMO frequency-selective channels and frequency-dependent I/Q imbalances. The proposed designs require much smaller pilot overhead than the existing designs, and also provide estimation mean-squares error optimality (under white noise) and general applicability to preamble as well as pilot-data-multiplexed symbols in MIMO systems with or without null guard tones. Performance analyses and simulation results corroborate advantages of the proposed designs.

We address the problem of label assignment in computer vision: given a novel 3D or 2D scene, we wish to assign a unique label to every site (voxel, pixel, superpixel, etc.). To this end, the Markov Random Field framework has proven to be a model of choice as it uses contextual information to yield improved classification results over locally independent classifiers. In this work we adapt a functional gradient approach for learning high-dimensional parameters of random fields in order to perform discrete, multi-label classification. With this approach we can learn robust models involving high-order interactions better than the previously used learning method. We validate the approach in the context of point cloud classification and improve the state of the art. In addition, we successfully demonstrate the generality of the approach on the challenging vision problem of recovering 3-D geometric surfaces from images.

Efficient pilot designs are vital as they improve system performance and resource utilization, especially in multi-input multi-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems. Most of the existing pilot designs do not consider in-phase and quadrature (I/Q) imbalance whose detrimental effects become more severe since the technology and applications are pushing towards smaller semiconductor node, smaller chip, cheaper equipment, and larger modulation order. This paper studies the impact of I/Q imbalance on the pilot designs, and presents several new pilot designs for equivalent channel estimation of MIMO OFDM systems with frequency-dependent I/Q imbalances. Advantages of the proposed designs are illustrated through simulation and analytical results in terms of estimation accuracy, error rate, overhead efficiency, and generality with the transmission formats (preamble-based or pilot-data-multiplexed systems, with or without null guard tones).

Contextual reasoning through graphical models such as Markov random fields often show superior performance against local classifiers in many domains. Unfortunately, this performance increase is often at the cost of time consuming, memory intensive learning and slow inference at testing time. Structured prediction for 3-D point cloud classification is one example of such an application. In this paper we present two contributions. First we show how efficient learning of a random field with higher-order cliques can be achieved using subgradient optimization. Second, we present a context approximation using random fields with high-order cliques designed to make this model usable online, onboard a mobile vehicle for environment modeling. We obtained results with the mobile vehicle on a variety of terrains, at 1/3 Hz for a map 25 times 50 meters and a vehicle speed of 1-2 m/s.

All practical OFDM systems insert null guard bands to facilitate easy filter implementation and emission spectrum control. The null guard bands introduce a constraint for optimal pilot designs in MIMO OFDM systems. Semiconductor downscaling and the use of high modulation order have signified the importance of pilot designs for OFDM systems in the presence of I/Q imbalances, but existing designs have not addressed the use of guard bands, the overhead efficiency, the estimation optimality, and general applicability in pilot data multiplexed format and/or MIMO setup. This paper proposes a pilot design for channel estimation of MIMO OFDM systems with guard bands and both transmit and receive frequency-dependent I/Q imbalances, which overcomes the limitations of the existing designs. Simulation and numerical results corroborate better performance, overhead advantage and applicability in pilot-data-multiplexed MIMO setup of the proposed design over the existing MIMO OFDM pilot design with I/Q imbalances.

The advances in semiconductor downscaling have empowered communication devices with significantly enhanced signal processing capability, but with the side effect of larger device impairments. Among those impairments, the I/Q imbalance is a critical one as it causes intercarrier interferences in orthogonal frequency division multiplexing (OFDM) systems. The I/Q imbalances in general exist at both transmitter and receiver, and they can be frequency-dependent. They may vary over time (e.g. due to sudden changes in temperature or the dynamics of the cooperating/relaying users). Under such scenarios, estimation and compensation of the combined effect of the channel and I/Q imbalances are crucial. This paper presents efficient pilot designs for such tasks. The advantages of the proposed designs over the existing ones in terms of estimation performance and overhead efficiency are corroborated by analysis and simulation results.