72 Hits in 1.9 sec

How Powerful Are Randomly Initialized Pointcloud Set Functions? [article]

Aditya Sanghi, Pradeep Kumar Jayaraman
2020 arXiv   pre-print
We study random embeddings produced by untrained neural set functions, and show that they are powerful representations which well capture the input features for downstream tasks such as classification, and are often linearly separable. We obtain surprising results that show that random set functions can often obtain close to or even better accuracy than fully trained models. We investigate factors that affect the representative power of such embeddings quantitatively and qualitatively.
arXiv:2003.05410v1 fatcat:ttfl35wahve3tcv2lxq4mk34my

Engineering Sketch Generation for Computer-Aided Design [article]

Karl D.D. Willis, Pradeep Kumar Jayaraman, Joseph G. Lambourne, Hang Chu, Yewen Pu
2021 arXiv   pre-print
Engineering sketches form the 2D basis of parametric Computer-Aided Design (CAD), the foremost modeling paradigm for manufactured objects. In this paper we tackle the problem of learning based engineering sketch generation as a first step towards synthesis and composition of parametric CAD models. We propose two generative models, CurveGen and TurtleGen, for engineering sketch generation. Both models generate curve primitives without the need for a sketch constraint solver and explicitly
more » ... r topology for downstream use with constraints and 3D CAD modeling operations. We find in our perceptual evaluation using human subjects that both CurveGen and TurtleGen produce more realistic engineering sketches when compared with the current state-of-the-art for engineering sketch generation.
arXiv:2104.09621v1 fatcat:xxm2z53t2nh7tfbryfa5kusyma

Quadtree Convolutional Neural Networks [chapter]

Pradeep Kumar Jayaraman, Jianhan Mei, Jianfei Cai, Jianmin Zheng
2018 Lecture Notes in Computer Science  
This paper presents a Quadtree Convolutional Neural Network (QCNN) for efficiently learning from image datasets representing sparse data such as handwriting, pen strokes, freehand sketches, etc. Instead of storing the sparse sketches in regular dense tensors, our method decomposes and represents the image as a linear quadtree that is only refined in the non-empty portions of the image. The actual image data corresponding to non-zero pixels is stored in the finest nodes of the quadtree.
more » ... on and pooling operations are restricted to the sparse pixels, leading to better efficiency in computation time as well as memory usage. Specifically, the computational and memory costs in QCNN grow linearly in the number of non-zero pixels, as opposed to traditional CNNs where the costs are quadratic in the number of pixels. This enables QCNN to learn from sparse images much faster and process high resolution images without the memory constraints faced by traditional CNNs. We study QCNN on four sparse image datasets for sketch classification and simplification tasks. The results show that QCNN can obtain comparable accuracy with large reduction in computational and memory costs.
doi:10.1007/978-3-030-01231-1_34 fatcat:gynykn7lc5bo5pntbs2qcgi34m

PointMask: Towards Interpretable and Bias-Resilient Point Cloud Processing [article]

Saeid Asgari Taghanaki, Kaveh Hassani, Pradeep Kumar Jayaraman, Amir Hosein Khasahmadi, Tonya Custis
2020 arXiv   pre-print
Deep classifiers tend to associate a few discriminative input variables with their objective function, which in turn, may hurt their generalization capabilities. To address this, one can design systematic experiments and/or inspect the models via interpretability methods. In this paper, we investigate both of these strategies on deep models operating on point clouds. We propose PointMask, a model-agnostic interpretable information-bottleneck approach for attribution in point cloud models.
more » ... ask encourages exploring the majority of variation factors in the input space while gradually converging to a general solution. More specifically, PointMask introduces a regularization term that minimizes the mutual information between the input and the latent features used to masks out irrelevant variables. We show that coupling a PointMask layer with an arbitrary model can discern the points in the input space which contribute the most to the prediction score, thereby leading to interpretability. Through designed bias experiments, we also show that thanks to its gradual masking feature, our proposed method is effective in handling data bias.
arXiv:2007.04525v1 fatcat:2amdw43p2nf2lou7iupu35w4o4

Submitral aneurysm: a rare cause of ventricular tachycardia

Pradeep Kumar, Jayaraman Balachander, Raja J Selvaraj
2012 Heart Asia  
A 42-year-old male presented with sustained ventricular tachycardia of left bundle branch block (LBBB) morphology with left axis deviation (figure 1A) that terminated with amiodarone infusion. He gave a history of episodic palpitations associated with giddiness. Transthoracic and transoesophageal echocardiography showed a wide necked submitral aneurysm measuring 5×5 cm with severe mitral regurgitation (figure 1B,C). Coronary angiogram revealed a dominant right coronary artery with spontaneous
more » ... ssection of mid and distal segments with splaying of the distal branches overlying the aneurysm (figure 2A,B) . Ventricular tachycardia of right and left bundle branch block morphologies was induced with single ventricular extrastimuli during electrophysiological testing. MRI clearly demonstrated the aneurysm below the mitral valve ( figure 2C,D) . The patient was prescribed oral amiodarone and advised surgical resection of submitral aneurysm with mitral valve repair or replacement. 1 DISCUSSION Submitral aneurysm is rare outside African blacks and uncommonly presents with ventricular arrhythmias. 2 While the aetiology is considered to be congenital, coronary abnormalities have occasionally been described in association with this condition, although the relationship of the coronary abnormality to the aneurysm is unclear in our patient.
doi:10.1136/heartasia-2012-010149 pmid:27326044 pmcid:PMC4832620 fatcat:em3ik4y7g5h67lng5prwkvw5iy

Computational interlocking furniture assembly

Chi-Wing Fu, Peng Song, Xiaoqi Yan, Lee Wei Yang, Pradeep Kumar Jayaraman, Daniel Cohen-Or
2015 ACM Transactions on Graphics  
Figure 1 : Some snapshots showing the assembly of MULTI-FUNCTION TABLE. Our method can plan a network of joints (e.g., Figure 2 ) that globally interlocks the component parts in the assembly; the input component parts are just simple 3D shapes without joint geometry. Abstract Furniture typically consists of assemblies of elongated and planar parts that are connected together by glue, nails, hinges, screws, or other means that do not encourage disassembly and re-assembly. An alternative approach
more » ... is to use an interlocking mechanism, where the component parts tightly interlock with one another. The challenge in designing such a network of interlocking joints is that local analysis is insufficient to guarantee global interlocking, and there is a huge number of joint combinations that require an enormous exploration effort to ensure global interlocking. In this paper, we present a computational solution to support the design of a network of interlocking joints that form a globally-interlocking furniture assembly. The key idea is to break the furniture complex into an overlapping set of small groups, where the parts in each group are immobilized by a local key, and adjacent groups are further locked with dependencies. The dependency among the groups saves the effort of exploring the immobilization of every subset of parts in the assembly, thus allowing the intensive interlocking computation to be localized within each small group. We demonstrate the effectiveness of our technique on many globally-interlocking furniture assemblies of various shapes and complexity.
doi:10.1145/2766892 fatcat:hlkidx2xfrf3tkbdojsfyzeiyu

UV-Net: Learning from Boundary Representations [article]

Pradeep Kumar Jayaraman, Aditya Sanghi, Joseph G. Lambourne, Karl D.D. Willis, Thomas Davies, Hooman Shayani, Nigel Morris
2021 arXiv   pre-print
We introduce UV-Net, a novel neural network architecture and representation designed to operate directly on Boundary representation (B-rep) data from 3D CAD models. The B-rep format is widely used in the design, simulation and manufacturing industries to enable sophisticated and precise CAD modeling operations. However, B-rep data presents some unique challenges when used with modern machine learning due to the complexity of the data structure and its support for both continuous non-Euclidean
more » ... ometric entities and discrete topological entities. In this paper, we propose a unified representation for B-rep data that exploits the U and V parameter domain of curves and surfaces to model geometry, and an adjacency graph to explicitly model topology. This leads to a unique and efficient network architecture, UV-Net, that couples image and graph convolutional neural networks in a compute and memory-efficient manner. To aid in future research we present a synthetic labelled B-rep dataset, SolidLetters, derived from human designed fonts with variations in both geometry and topology. Finally we demonstrate that UV-Net can generalize to supervised and unsupervised tasks on five datasets, while outperforming alternate 3D shape representations such as point clouds, voxels, and meshes.
arXiv:2006.10211v2 fatcat:vccqocfodvcihke2kk4lx6mvli

Interactive Line Drawing Recognition and Vectorization with Commodity Camera

Pradeep Kumar Jayaraman, Chi-Wing Fu
2014 Proceedings of the ACM International Conference on Multimedia - MM '14  
This paper presents a novel method that interactively recognizes and vectorizes hand-drawn strokes in front of a commodity webcam. Compared to existing methods, which recognize strokes on a completed drawing, our method captures both spatial and temporal information of the strokes, and faithfully vectorizes them with timestamps. By this, we can avoid various stroke recognition ambiguities, enhance the vectorization quality, and recover the stroke drawing order. This is a challenging problem,
more » ... uiring robust tracking of pencil tip, accurate modeling of pen-paper contact, handling pen-paper and hand-paper occlusion, while achieving interactive performance. To address these issues, we develop the following novel techniques. First, we perform robust spatiotemporal tracking of pencil tip by extracting discriminable features, which can be classified with a fast cascade of classifiers. Second, we model the pen-paper contact by analyzing the edge-profile of the acquired trajectory and extracting the portions related to individual strokes. Lastly, we propose a spatio-temporal method to reconstruct meaningful strokes, which are coherent to the stroke drawing continuity and drawing order. By integrating these techniques, our method can support interactive recognition and vectorization of drawn strokes that are faithful to the actual strokes drawn by the user, and facilitate the development of various multimedia applications such as video scribing, cartoon production, and pen input interface.
doi:10.1145/2647868.2654939 dblp:conf/mm/JayaramanF14 fatcat:kjnon4pfejagbozhxn2jt5dh5m

Pace mapping in the atrium using bipolar electrograms from widely spaced electrodes

Raja J. Selvaraj, Sreekanth Yerram, Pradeep Kumar, Santhosh Satheesh, Ajith Ananthakrishna Pillai, Mahesh Kumar Saktheeswaran, Jayaraman Balachander
2015 Journal of Arrhythmia  
Background: Pace mapping is a useful tool but is of limited utility for the atrium because of poor spatial resolution. We investigated the use of bipolar electrograms recorded from widely spaced electrodes in order to improve the resolution of pace mapping. Methods: This prospective study included patients undergoing a clinical electrophysiology study. Unipolar pacing from either the superior or inferior lateral right atrium was performed to simulate atrial tachycardia. Twelve-lead
more » ... grams were recorded during pacing as a template. In addition, three intracardiac bipolar electrograms from a set of widely spaced electrodes were also recorded. Subsequently, unipolar pacing was performed from electrodes at known distances from the initial pacing site, and the morphology of P waves in the electrocardiogram and bipolar electrograms were compared with that of the template. Morphological comparison was performed by a cardiologist and by automated computerized matching. Spatial resolution was calculated as the minimum distance at which there was no match. Results: Fifteen patients participated in the study. Distance at which differences in morphology were noted was smaller in the bipolar electrograms compared to that indicated by P waves in the electrocardiogram, when matched by the cardiologist (6.1 73.8 mm vs. 9.97 5.2 mm, p¼0.012) or by automated analysis (4 70 mm vs. 9.97 4 mm, po 0.001). Conclusions: Use of three bipolar electrograms recorded from a set of widely spaced electrodes in the right atrium improves the resolution of pace mapping compared to that using P waves from surface electrocardiograms alone.
doi:10.1016/j.joa.2015.02.002 pmid:26550082 pmcid:PMC4600891 fatcat:w6gcctotcrflhgyms2uao6o62m

BRepNet: A topological message passing system for solid models [article]

Joseph G. Lambourne, Karl D.D. Willis, Pradeep Kumar Jayaraman, Aditya Sanghi, Peter Meltzer, Hooman Shayani
2021 arXiv   pre-print
Jayaraman et al. [20] uses convolution layers to create input features from grids of 3D points and normal vectors, while Cao et al.  ... 
arXiv:2104.00706v2 fatcat:q2em4c5mafcjnal3d72wqugmru

JoinABLe: Learning Bottom-up Assembly of Parametric CAD Joints [article]

Karl D.D. Willis, Pradeep Kumar Jayaraman, Hang Chu, Yunsheng Tian, Yifei Li, Daniele Grandi, Aditya Sanghi, Linh Tran, Joseph G. Lambourne, Armando Solar-Lezama, Wojciech Matusik
2022 arXiv   pre-print
Physical products are often complex assemblies combining a multitude of 3D parts modeled in computer-aided design (CAD) software. CAD designers build up these assemblies by aligning individual parts to one another using constraints called joints. In this paper we introduce JoinABLe, a learning-based method that assembles parts together to form joints. JoinABLe uses the weak supervision available in standard parametric CAD files without the help of object class labels or human guidance. Our
more » ... ts show that by making network predictions over a graph representation of solid models we can outperform multiple baseline methods with an accuracy (79.53%) that approaches human performance (80%). Finally, to support future research we release the Fusion 360 Gallery assembly dataset, containing assemblies with rich information on joints, contact surfaces, holes, and the underlying assembly graph structure.
arXiv:2111.12772v2 fatcat:w44jtwy2xbgltfzmxoc6k3jity

Ultrasound Guided Longitudinal Supra-Inguinal Fascia Iliaca Block for Hip Surgeries- A Prospective Study

Rajasekhar Thondamanati, Subhashree Jayaraman, Madhusalini Kondapalli, Pradeep Kumar Koramutla
2020 Journal of Evidence Based Medicine and Healthcare  
A BS T R A C T BACKGROUND Hip fractures are very common in geriatric population with incidence increasing every year because of increased life expectancy. Effective postoperative pain management can result in early ambulation and reduced hospital stay. Peripheral nerve block techniques have been widely used for providing postoperative analgesia because of their less systemic effects among which, fascia iliaca compartment block will provide complete analgesia in hip surgeries as it blocks both
more » ... moral and lateral femoral cutaneous nerves. METHODS We conducted this case series to evaluate the analgesic efficacy of ultrasound guided suprainguinal fascia iliaca block with 30 ml of 0.25% bupivacaine with 0.5 µg/Kg dexmedetomidine. After obtaining institutional ethics committee and written informed consent, we included 50 patients belonging to ASA PS1, 2 posted for various hip surgeries. Patients who did not give consent and patients with history of coagulopathy were excluded from the study. All patients underwent the procedure under spinal anaesthesia. Once the procedure was over, patient was shifted to postoperative ICU, under sterile aseptic precautions, and ultrasound guided suprainguinal fascia iliaca block was given. Postoperative analgesia was assessed using VAS scale and patient's satisfaction using Likert's satisfaction score at 20 minutes, 6th hour, 12th hour, and 24 th hour. Rescue analgesia was given with tramadol 0.5-1 mg/Kg when the VAS score was more than 4 and the total dose given in the 24 hours was noted. RESULTS We found that VAS score (ANOVA test) was less at 20 minutes, 6 th hour and 12 th hour (p < 0.001), which was statistically significant, and less rescue analgesic was required. Hence, we concluded that suprainguinal fascia iliaca block provides effective analgesia for hip procedures in elderly patients. CONCLUSIONS Suprainguinal fascia iliaca compartment block with 30 ml of 0.25% bupivacaine with 0.5 µg/Kg dexmedetomidine provides effective and prolonged analgesia with less rescue analgesic requirement and less systemic effects.
doi:10.18410/jebmh/2020/301 fatcat:qbucffw4fvg5tkuv3s34hywkpi

Comparative analysis of subgingival red complex bacteria in obese and normal weight subjects with and without chronic periodontitis

Jaideep Mahendra, Snophia Suresh, AngabakkamRajasekaran Pradeep Kumar, Gurdeep Singh, Selvaraj Jayaraman, Roshini Paul
2017 Journal of Indian Society of Periodontology  
Obesity is one of the systemic conditions which influence the onset and progression of periodontal disease and it is stated that the metabolic changes associated with obesity may contribute to alteration in subgingival microbial flora. Our study was aimed to quantify and compare the red complex microorganisms in obese or overweight and normal weight participants with and without chronic periodontitis to identify obesity as a risk for the presence of red complex bacteria. The study group
more » ... d of 120 participants of age between 20 and 45 years of both the sexes. According to periodontal status, the participants were categorized into four groups as follows: thirty overweight or obese individuals with generalized chronic periodontitis (Group I), thirty normal weight individuals with chronic periodontitis (Group II), thirty overweight or obese individuals with healthy periodontium (Group III), and thirty normal weight individuals with healthy periodontium (Group IV). After the assessment of periodontal parameters, subgingival plaque sample collection was carried out to quantify the red complex bacteria by real-time polymerase chain reaction. Increase in red complex bacterial count was seen in group I compared to other groups. A positive correlation of red complex bacteria with body mass index and waist circumference was seen in Group I and III. In our study, obese individuals with periodontal disease harbored increased red complex bacteria. This states that the obesity could be a risk for the colonization of red complex microorganisms, which in turn may further lead to periodontal inflammation.
doi:10.4103/jisp.jisp_241_17 pmid:29440783 pmcid:PMC5803872 fatcat:pvqdt6sornc63mjosksnwu5vta

UVStyle-Net: Unsupervised Few-shot Learning of 3D Style Similarity Measure for B-Reps [article]

Peter Meltzer, Hooman Shayani, Amir Khasahmadi, Pradeep Kumar Jayaraman, Aditya Sanghi, Joseph Lambourne
2021 arXiv   pre-print
The benefits of B-Reps over discrete representations are demonstrated in Jayaraman et al.  ...  [9] and Jayaraman et al. [16] , we perform pre-training using 26 classes (combining upper and lower case examples).  ... 
arXiv:2105.02961v3 fatcat:sxdznzourzddnbtvaz3g4yy2kq

RobustPointSet: A Dataset for Benchmarking Robustness of Point Cloud Classifiers [article]

Saeid Asgari Taghanaki, Jieliang Luo, Ran Zhang, Ye Wang, Pradeep Kumar Jayaraman, Krishna Murthy Jatavallabhula
2021 arXiv   pre-print
The 3D deep learning community has seen significant strides in pointcloud processing over the last few years. However, the datasets on which deep models have been trained have largely remained the same. Most datasets comprise clean, clutter-free pointclouds canonicalized for pose. Models trained on these datasets fail in uninterpretible and unintuitive ways when presented with data that contains transformations "unseen" at train time. While data augmentation enables models to be robust to
more » ... ously seen" input transformations, 1) we show that this does not work for unseen transformations during inference, and 2) data augmentation makes it difficult to analyze a model's inherent robustness to transformations. To this end, we create a publicly available dataset for robustness analysis of point cloud classification models (independent of data augmentation) to input transformations, called RobustPointSet. Our experiments indicate that despite all the progress in the point cloud classification, there is no single architecture that consistently performs better -- several fail drastically -- when evaluated on transformed test sets. We also find that robustness to unseen transformations cannot be brought about merely by extensive data augmentation. RobustPointSet can be accessed through
arXiv:2011.11572v5 fatcat:n4bb5tmbk5czhmehpad4glpwsu
« Previous Showing results 1 — 15 out of 72 results