Meta tags:
author=Leonid Keselman;

Headings (most frequently used words):

for, based, depth, tracking, intel, with, of, models, dynamics, low, projects, differentiable, rendering, 3d, optimizing, realsense, stereoscopic, hand, power, networks, stereo, gina, leonid, keselman, research, other, flexible, techniques, gaussians, algorithms, from, pairwise, user, preferences, discovering, multiple, algorithm, configurations, fuzzy, metaballs, approximate, algebraic, surfaces, venue, analytics, simple, alternative, to, citation, metrics, direct, fitting, gaussian, mixture, cameras, rigid, body, articulated, mesh, 400, compact, vcsel, projector, image, enhancement, real, time, box, measurement, dashpoint, cost, human, interface, device, reconstruction, probabilistic, pixel, correspondence, search, dice, stacking, dynamic, manipulation, task, introspective, neural, stochastic, sampling, parametric, policies, physical, simulation, maze, bot, learning, implicit, communication, strategies, computational, text, summarization, superresolution, micrscopy, automatically, building, restaurant, ontologies, beyond, correlation, the, financial, market, gradient, learned, matching, multimodal, natural, language, inference, cnns, model, classification, wide, angle, lenses, doctor, bayes, level, set, and, segmentation, dequantization, data, golf, swing, monitoring, project, tetra, collaborative, robot, state, estimation, design, wireless, sensor, platform,

Text of the page (most frequently used words):
and (48), for (28), the (27), code (16), with (15), intel (14), paper (13), stanford (13), using (12), leonid (12), #keselman (12), depth (10), computer (10), arxiv (9), from (8), website (8), time (8), 2015 (8), tracking (8), 2016 (8), stereo (8), learning (8), vision (8), work (7), algorithm (7), poster (7), video (7), 2017 (7), patent (7), realsense (7), low (6), real (6), data (6), some (6), martial (6), hebert (6), sensors (5), robot (5), results (5), networks (5), models (5), optimization (5), our (5), conference (5), design (4), berkeley (4), sensor (4), show (4), cameras (4), systems (4), used (4), other (4), based (4), computational (4), classification (4), visual (4), recognition (4), natural (4), language (4), can (4), algorithms (4), slides (4), robotics (4), 2018 (4), cmu (4), method (4), rgb (4), dynamics (4), differentiable (4), gina (3), wireless (3), platform (3), 2011 (3), project (3), state (3), estimation (3), that (3), cost (3), about (3), stereoscopic (3), neural (3), how (3), reconstruction (3), solve (3), dynamic (3), simple (3), manipulation (3), include (3), projects (3), research (3), 2013 (3), projector (3), stan (3), melax (3), sterling (3), orsten (3), graphics (3), cvpr (3), details (3), mesh (3), science (3), approximate (3), rendering (3), 2023 (3), where (3), worked (3), 2010 (2), helped (2), these (2), small (2), testing (2), navigation (2), power (2), multiple (2), collaborative (2), platforms (2), communication (2), golf (2), swing (2), producing (2), such (2), detection (2), cells (2), level (2), top (2), accuracy (2), analysis (2), imaging (2), shape (2), into (2), image (2), pre (2), trained (2), network (2), single (2), cs231n (2), model (2), inference (2), tasks (2), correspondence (2), matching (2), gradient (2), track (2), financial (2), social (2), information (2), restaurant (2), dimensions (2), text (2), summarization (2), reinforcement (2), implicit (2), strategies (2), aaron (2), maze (2), 745 (2), optimizing (2), very (2), sampling (2), via (2), designed (2), task (2), 2012 (2), fast (2), maps (2), human (2), box (2), system (2), compact (2), was (2), acm (2), hand (2), solver (2), articulated (2), objects (2), this (2), rigid (2), body (2), machine (2), hands (2), workshops (2), r200 (2), d400 (2), 2019 (2), fitting (2), gaussian (2), mixture (2), produces (2), more (2), all (2), citation (2), venue (2), renderer (2), fuzzy (2), metaballs (2), youtube (2), discovering (2), configurations (2), international (2), efficient (2), user (2), preferences (2), gaussians (2), have (2), assistant (2), fei (2), university (2), part (2), including (2), phd (2), source, jon, barron, anita, flynn, thomas, watteyne, build, wrote, firmware, validating, functionality, guidance, inertial, mote, gram, humphrey, ryan, julian, eric, yuan, efficacy, wiimote, mobile, ee149, embedded, tetra, built, demonstration, weight, monitor, ankur, mehta, monitoring, smooth, normals, quantized, kinect, dequantization, implemented, deformable, pipeline, red, blood, cs279, structure, organization, biomolecules, set, segmentation, detecting, disease, short, description, symptoms, obtained, nearly, cs221, artificial, intelligence, doctor, bayes, introduce, various, projection, functions, cs448i, display, wide, angle, lenses, embedding, function, imagenet, got, art, view, shapenet40, convolutional, cnns, explored, augmented, cs224u, understanding, multimodal, implementations, standard, along, experiments, cs231a, learned, graph, development, over, 20th, century, cs224w, beyond, correlation, market, reviews, semantics, relationships, between, cuisines, businesses, properties, useful, recommendations, yelp, dataset, cs270, modeling, biomedical, automatically, building, ontologies, implementation, faster, storm, compressed, sensing, cs371, biology, four, superresolution, micrscopy, simplified, encoders, stages, ludwig, schubert, cs224n, processing, discover, collusion, context, iterated, prisoner, dilemma, goodman, cs234, deep, making, dof, puma, arm, cs225a, experimental, bot, alex, spitzer, optimizers, match, observed, trajectory, chris, atkeson, physical, simulation, environments, stochastic, parametric, policies, improve, fine, grained, style, transfer, 824, introspective, hunter, goforth, solved, imitation, 741, mechanics, dice, stacking, coursework, side, unpublished, performing, probabilistic, pixel, search, finger, microcontroller, optics, tricks, hci, ideas, dashpoint, interface, device, cuboids, accurate, their, even, bin, packing, measurement, filter, enhance, clean, streams, enhancement, dense, configurable, vcsel, responsibilities, included, performance, components, asic, contributions, laser, pattern, 400, besides, publications, above, publicly, disclosed, consumer, laptop, won, best, award, siggraph, symposium, interactive, games, physics, engine, interfaces, invited, talk, workshop, includes, further, efficiency, tools, annotation, process, john, iselin, woodfill, anders, grunnet, jepsen, achintya, bhowmik, technical, series, displays, formulation, gmms, directly, geometric, triangular, robust, improvement, registration, both, meshes, frames, direct, bibliometric, scientometric, main, two, are, assigning, value, venues, organize, them, distinct, subfields, without, resulting, scores, rank, universities, scholarly, output, produce, responsive, author, metric, ieee, joint, digital, libraries, analytics, alternative, metrics, 2022, interpretable, representation, which, call, focuses, shapes, silhouettes, sacrifices, fidelity, utility, runtimes, high, quality, european, oral, algebraic, surfaces, benefits, ensemble, given, during, course, planning, odometry, automation, katherine, shih, steinfeld, perform, black, configuration, policy, intelligent, robots, pairwise, expanded, optical, flow, colored, exports, flexible, techniques, interested, linkedin, google, scholar, github, email, focus, teaching, eecs, lab, kris, pister, cs131, silvio, savarese, primarily, hardware, asics, additionally, develop, demos, trade, shows, ces, interaction, devices, active, illumination, software, apis, student, advisor, carnegie, mellon, school, institute,

Text of the page (random words):
leonid keselman leonid keselman i am a phd student at the robotics institute part of the school of computer science at carnegie mellon university where i work on 3d computer vision my phd advisor is martial hebert from 2011 to 2017 i worked at intel as part of intel realsense i primarily designed computer vision algorithms for efficient hardware asics including the intel realsense r200 and d400 rgb d sensors additionally i worked on software apis active illumination systems human computer interaction devices and helped develop demos for trade shows including ces 2012 2016 i have an ms in computer science ai focus from stanford university where i was a research assistant for silvio savarese and a teaching assistant for fei fei li cs131 cs231n i have a bs in eecs from uc berkeley where i worked in kris pister s lab email github google scholar linkedin research i m interested in computer vision machine learning optimization graphics and robotics flexible techniques for differentiable rendering with 3d gaussians leonid keselman martial hebert arxiv 2023 arxiv code website we show how shape reconstruction with 3d gaussians can be expanded to include differentiable optical flow colored mesh exports and more optimizing algorithms from pairwise user preferences leonid keselman katherine shih martial hebert aaron steinfeld international conference on intelligent robots and systems 2023 arxiv code website we show how to perform efficient black box optimization of algorithm configuration from user preferences results include intel realsense stereo cameras and a robot social navigation policy discovering multiple algorithm configurations leonid keselman martial hebert international conference on robotics and automation 2023 arxiv code website youtube we show the benefits of discovering an ensemble of configurations for a given algorithm during the course of optimization results on stereo planning and visual odometry fuzzy metaballs approximate differentiable rendering with algebraic surfaces leonid keselman martial hebert european conference on computer vision oral 2022 arxiv code website youtube an approximate differentiable renderer for a compact interpretable representation which we call fuzzy metaballs our approximate renderer focuses on rendering shapes via depth maps and silhouettes it sacrifices fidelity for utility producing fast runtimes and high quality gradient information that can be used to solve vision tasks venue analytics a simple alternative to citation based metrics leonid keselman acm ieee joint conference on digital libraries 2019 arxiv code slides website a bibliometric scientometric project our main two results are a method of assigning value to all venues in computer science and a method to organize them into distinct subfields all without using citation data the resulting venue scores can be used to rank universities by scholarly output and produce a responsive author level metric direct fitting of gaussian mixture models leonid keselman martial hebert computer and robot vision conference 2019 arxiv code slides website a formulation for fitting gaussian mixture models gmms directly to geometric objects such as a triangular mesh this method produces more robust results and produces an improvement in 3d registration for both meshes and rgb d frames intel realsense stereoscopic depth cameras leonid keselman john iselin woodfill anders grunnet jepsen achintya bhowmik cvpr workshops computational cameras and displays 2017 arxiv technical and design details of the intel realsense r200 and d400 series rigid body dynamics for articulated mesh tracking leonid keselman sterling orsten stan melax cvpr workshops hands 2015 slides an invited talk for the hands 2015 workshop at cvpr 2015 this includes further details about the efficiency of our rigid body solver our machine learning tools and some details about our data annotation process dynamics based hand tracking stan melax leonid keselman sterling orsten graphics interfaces 2013 arxiv code using a physics engine e g a dynamics solver to track 3d articulated objects in real time dynamics based hand tracking stan melax leonid keselman sterling orsten acm siggraph symposium on interactive 3d graphics and games 2013 poster a tracking algorithm that was real time on a consumer laptop won best poster award intel projects besides my work on the realsense depth sensors and the publications above a sampling of my publicly disclosed work intel realsense 400 intel 2016 08 15 my responsibilities included system performance components of the stereo algorithm on the imaging asic and contributions to the design of laser projector pattern compact vcsel projector intel 2016 06 27 patent patent 2 patent 3 a low cost dense configurable projector system for rgb d depth sensors depth image enhancement intel 2015 08 06 patent algorithms to filter enhance and clean up rgb d data streams real time box measurement intel 2015 04 08 video video 2 using a single depth sensor real time detection of cuboids accurate estimation of their dimensions and even some bin packing dashpoint a low cost low power human interface device intel 2013 06 07 patent patent 2 finger tracking on a microcontroller with optics tricks and some hci ideas stereoscopic depth reconstruction with probabilistic pixel correspondence search intel 2012 07 24 patent a fast method for performing stereo depth maps other projects these include coursework side projects and unpublished research work dice stacking a dynamic manipulation task cmu 16 741 mechanics of manipulation 2018 12 05 paper video code with hunter goforth we designed a manipulation task and solved it with imitation learning introspective neural networks cmu 16 824 visual learning and recognition 2018 05 15 paper using pre trained neural networks to improve fine grained recognition via style transfer stochastic sampling of parametric policies cmu 16 745 dynamic optimization 2018 05 05 paper using a very simple algorithm to solve some very simple environments optimizing for physical simulation cmu 16 745 dynamic optimization 2018 03 22 code with chris atkeson and alex spitzer using optimizers to match an observed trajectory a maze bot stanford cs225a experimental robotics 2017 06 12 paper video video 2 making a 6 dof puma arm solve a maze with real time vision and tracking learning implicit communication strategies stanford cs234 deep reinforcement learning 2017 06 10 work with aaron goodman on used reinforcement learning to discover implicit collusion strategies in the context of an iterated prisoner s dilemma computational models for text summarization stanford cs224n natural language processing 2017 03 18 paper video code poster work with ludwig schubert on simplified encoders stages for text summarization superresolution micrscopy stanford cs371 computational biology in four dimensions 2017 03 16 code slides an implementation of faster storm using compressed sensing automatically building restaurant ontologies stanford cs270 modeling biomedical systems 2017 03 15 paper poster using the yelp dataset of reviews to model the semantics and relationships between cuisines businesses and other properties useful for restaurant recommendations beyond correlation networks for the financial market stanford cs224w social and information network analysis 2016 12 07 paper using graph models we track the development of financial networks over the 20th century gradient learned models for stereo matching stanford cs231a computer vision from 3d reconstruction to recognition 2016 06 07 paper code some re implementations of standard stereo correspondence algorithms along with experiments using classification for stereo matching multimodal natural language inference stanford cs224u natural language understanding 2016 06 06 paper video we explored how natural language inference tasks can be augmented with visual data cnns for 3d model classification stanford cs231n convolutional neural networks for visual recognition 2016 03 08 paper poster 3d shape classification by learning an embedding function into a 2d image and using a pre trained imagenet network at the time got state of the art results for single view classification on shapenet40 wide angle stereo lenses stanford cs448i computational imaging and display 2016 03 07 paper poster we introduce various projection functions in the analysis of stereoscopic depth sensors doctor bayes stanford cs221 artificial intelligence 2015 12 12 website paper code poster detecting disease from a short description of symptoms in some small testing obtained nearly 90 top 5 accuracy and about 60 top 1 accuracy level set based tracking and segmentation stanford cs279 structure and organization of biomolecules and cells 2015 12 04 paper code we implemented a detection and deformable tracking pipeline for red blood cells dequantization of depth data other 2015 04 22 code an o 1 time algorithm for producing smooth normals for quantized data such as the kinect golf swing monitoring other 2011 07 21 work with ankur mehta built a demonstration platform that used wireless low weight low cost sensor platforms to monitor a golf swing project tetra collaborative robot state estimation uc berkeley ee149 embedded systems 2011 07 21 with humphrey hu ryan julian and eric yuan a project to show the efficacy of multiple robot collaborative state estimation using wiimote cameras mobile robot platforms and real time wireless communication gina low power design uc berkeley 2010 08 22 for testing and validating the functionality of the gina guidance and inertial navigation mote a 1 6 gram sensor platform gina wireless sensor platform uc berkeley 2010 06 22 i helped anita flynn and thomas watteyne build these small sensors and wrote firmware design and source code from jon barron s website