Abstract

Accurate localization of other traffic participants is a vital task in autonomous driving systems. State-of-the-art systems employ a combination of sensing modalities such as RGB cameras and LiDARs for localizing traffic participants, but monocular localization demonstrations have been confined to plain roads. We demonstrate — to the best of our knowledge — the first results for monocular object localization and shape estimation on surfaces that are non-coplanar with the moving ego vehicle mounted with a monocular camera. We approximate road surfaces by local planar patches and use semantic cues from vehicles in the scene to initialize a local bundle-adjustment like procedure that simultaneously estimates the 3D pose and shape of the vehicles, and the orientation of the local ground plane on which the vehicle stands. We also demonstrate that our approach transfers from synthetic to real data, without any hyperparameter-/fine-tuning. We evaluate the proposed approach on the KITTI and SYNTHIA-SF benchmarks, for a variety of road plane configurations. The proposed approach significantly improves the state-of-the-art

Abstract

Motion planning for manipulators under task space constraints is difficult as it constrains the joint config- urations to always lie on an implicitly defined manifold. It is possible to view task constrained motion planning as an opti- mization problem with non-linear equality constraints, which can be solved by general non-linear optimization techniques. In this paper, we present a novel custom optimizer which exploits the underlying structure present in many task constraints. At the core of our approach are some simple reformulations, which when coupled with the method of alternating projection, leads to an efficient convex optimization based routine for computing a feasible solution to the task constraints. We subsequently build on this result and use the concept of Augmented Lagrangian to guide the feasible solutions towards those that also minimize the user defined cost function. We show that the proposed optimizer is fully distributive and thus, can be easily parallelized. We validate our formulation on some common robotic benchmark problems. In particular, we show that the proposed optimizer achieves cyclic motion in the joint space corresponding to a similar nature trajectory in the task space. Furthermore, as a baseline, we compare the proposed optimizer with an off-the-shelf non-linear solver provide in open source package SciPy. We show that for similar task constraint residuals and smoothness cost, it can be upto more than three times faster than the SciPy alternative

Abstract

Recently there have been a large number of studies on embedding large-scale information networks using low-dimensional, neighborhood and community aware node representations. Though the performance of these embedding models have been better than traditional methods for graph mining applications, little is known about what these representations encode, or why a particular node representation works better for certain tasks. Our work presented here constitutes the first step in decoding the black-box of vector embeddings of nodes by evaluating their effectiveness in encoding elementary properties of a node such as page rank, degree, closeness centrality, clustering coefficient, etc. We believe that a node representation is effective for an application only if it encodes the application-specific elementary properties of nodes. To unpack the elementary properties encoded in a node representation, we evaluate the representations on the accuracy with which they can model each of these properties. Our extensive study of three state-of-the-art node representation models (DeepWalk, node2vec and LINE) on four different tasks and six diverse graphs reveal that node2vec and LINE best encode the network properties of sparse and dense graphs respectively. We correlate the model performance obtained for elementary property prediction tasks with the high-level downstream applications such as link prediction and node classification, and visualize the task performance vector of each model to understand the semantic similarity between the embeddings learned by various models. Our first study of the node embedding models for outlier detection reveals that node2vec and DeepWalk identify outliers well for sparse and dense graphs respectively. Our analysis highlights that the proposed elementary property prediction tasks help in unearthing the important features responsible for the given node embedding model to perform well for a given downstream task. This understanding would facilitate in picking the right model for a given downstream task.

Abstract

We use automatic speech recognition to assess spoken English learner pronunciation based on the authentic intelligibility of the learners' spoken responses determined from support vector machine (SVM) classifier or deep learning neural network model predictions of transcription correctness. Using numeric features produced by PocketSphinx alignment mode and many recognition passes searching for the substitution and deletion of each expected phoneme and insertion of unexpected phonemes in sequence, the SVM models achieve 82% agreement with the accuracy of Amazon Mechanical Turk crowdworker transcriptions, up from 75% reported by multiple independent researchers. Using such features with SVM classifier probability prediction models can help computer-aided pronunciation teaching (CAPT) systems provide intelligibility remediation.

Abstract

Diabetic retinopathy (DR) is one of the most common reasons for blindness in the working-age population of world. Diabetic Retinopathy is an eye disease, which occurs with long-standing untreated diabetes. Progression to vision impairment can be slowed down or stopped if DR is detected on time; In detection or screening of DR, automatic methods can play an important role. In this paper, we proposed a novel method to detect hard exudates with high accuracy with respect to lesion level. In the present method we initially detected the possible candidate exudate lesions by using the back ground subtraction methodology. Following the subsequent steps, in the last stage of algorithm we removed the false exudate lesion detections using the de-correlation stretch based method. We tested our algorithm on publicly available DiaretDB database, which contains the ground truth for all images. We achieved high performance results such as sensitivity of 0.87 and F-Score of 0.78 and Positive Predict Value (PPV) of 0.76 for hard exudate lesion level detection, compared to the existing state of art techniques.

Abstract

Forest fires are the most frequent phenomenon during the summer season in India, and especially in the hilly terrains of Uttarakhand forests. Remote sensing sensors such as Moderate Resolution Imaging Spectroradiometer (MODIS), Advanced Very High Resolution Radiometer (AVHRR), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and Visible Infrared Imaging Radiometer Suite (VIIRS) with coarse spatial resolution on board different satellites were used to detect the forest fires across the world. Landsat-8 Operational Land Imager (OLI) data has the better spatial resolution (30m) as compared with the MODIS and VIIRS, therefore useful to detect the smaller fires. Nainital district in Uttarakhand state was severely affected by the massive forest fire events occurred during April-May, 2016. The main objective of the study is to identify the potential of Landsat-8 data in detecting the forest fire for varying topographic region like Nainital. Landsat-8 data acquired on 28 th April 2016 and 1 st May 2016 has been used in this study. The results obtained from Landsat-8 data are compared with the MODIS fire products and showed an improvement in the detection of small fires.

Abstract

We propose flexible-architecture for soft-input soft-output (SISO) decoder with radix-2/4/8 modes to support multiple data-rates. This work presents designs of major internal blocks of SISO decoder using extensive steering logic to support multiple radix operating-modes. These architectures enable efficient clock-gating of our decoder for low-power consumption in different operating modes. Subsequently, we have aggregated eight SISO decoders with quadratic-permutation-polynomial (QPP) interleavers/de-interleavers to design parallel-turbo decoder architecture which can operate in multi-radix mode. Suggested SISO decoder is ASIC synthesized and post-layout simulated in UMC 65 nm-CMOS process. Performance analyses in AWGN channel environment showed that the bit-error-rate (BER) of 10 -4 could be achieved at 5 dB and 0.8 dB for SISO and turbo decoders respectively. Implementation result shows that the suggested SISO decoder could achieve throughput in the range 270-810 Mbps with the corresponding power consumption range of 12.24-37.67 mW. In comparison to the state-of-the-art, our design achieved 38% higher throughput and 61% lower power consumption. Similarly, our multi-radix parallel-turbo decoder is hardware prototyped in 28 nm-CMOS Zynq-FPGA board. It delivers a range of data-rates from 80-320 Mbps operating at 160 MHz of clock frequency for 8 iterations.

Abstract

Emotion evoked by an advertisement plays a key role in influencing brand recall and eventual consumer choices. Automatic ad affect recognition has several useful applications. However, the use of content-based feature representations does not give insights into how affect is modulated by aspects such as the ad scene setting, salient object attributes and their interactions. Neither do such ap- proaches inform us on how humans prioritize visual information for ad understanding. Our work addresses these lacunae by decom- posing video content into detected objects, coarse scene structure, object statistics and actively attended objects identified via eye- gaze. We measure the importance of each of these information channels by systematically incorporating related information into ad affect prediction models. Contrary to the popular notion that ad affect hinges on the narrative and the clever use of linguistic and social cues, we find that actively attended objects and the coarse scene structure better encode affective information as compared to individual scene objects or conspicuous background elements.

Abstract

Bug report filing is a major part of software maintenance. Due to the asynchronous nature of the bug filing process, duplicate bug reports are filed. Detecting duplicate bug reports is an important aspect of software maintenance since the same bug should not be assigned to different developers. In this poster, we present Deep Word Embedding Network for computing similarity between two bug reports for the task of duplicate bug report detection. We propose to learn a two step model to calculate similarity between two bug reports by means of word embeddings and a deep neural network. We run experiments on two large datasets of Mozilla Project and Open Office Project and compare the proposed approach with baselines and related approaches. Through this initial work, we show that a combination of word embeddings and deep neural networks can be used to improve duplicate bug report detection.

Abstract

Several designer controllable parameters and their effect on various specifications like switching time, sampling speed and input amplitude of all-optical Binary Delta Sigma Modulator (BDSM) are analyzed and possibility of enhancing the performance is discussed.