225 research outputs found
Performance analysis of the WiNC2R platform:
No full textA Cognitive Radio (CR) is an intelligent transceiver device, able to support multiple technologies, dynamic re-configurability, ease of programming and collaboration with other CR devices to improve the communication efficiency. The two key requirements for an efficient CR implementation are flexibility in operation/programming and speed.
WiNC2R (Winlab Network Centric Cognitive Radio) achieves high speed of operation using its hardware platform and flexibility using its software-configurable architecture. The current WiNC2R architecture implements an 802.11a-like OFDM flow. We evaluate the WiNC2R hardware architecture to see the modularity in the architecture, separation of data and control flow and the performance in terms of latency and throughput. To test the system, the Xilinx Bus Functional Model environment, which is designed to test the IBM standard bus-architecture-based hardware systems, is used. We use a simple ALOHA protocol in the MAC layer to communicate between two WiNC2R nodes and evaluate the performance under the best-case scenario, where the performance is only hindered by the architecture itself rather than external conditions like channel state.
The results of our basic experiments showed that for a single OFDM 802.11a-like flow, the Unit Control Modules (UCM) were idle for almost 80% of the total processing time.
We then tested the WiNC2R system to study the effects of changing the frame size. It was seen that the latencies in the WiNC2R transmitter are frame-size dependent while those in the receiver mainly depend on the size of the data in the last chunk rather than the size of the whole frame. We suggest that chunk size should be 2 OFDM symbols, and chunking be moved to MAC layer for better performance. We give analytical estimates of resulting performance improvement. In the next experiment, we describe virtualization in the WiNC2R by adding more flows. We describe the steps to implement the additional flows and estimate maximum number of concurrent flows possible.
In the last analysis, we show the effect of operating clock frequency on the performance. We prove that at 250 MHz operating frequency and 2 OFDM symbols per chunk, the current WiNC2R implementation will be able to satisfy the SIFS criterion.M.S.Includes bibliographical references (p. 72-73)by Sumit Satarka
Sparse Methods for Robust and Efficient Visual Recognition
No full textVisual recognition has been a subject of extensive research in computer vision. A vast literature exists on feature extraction and learning methods for recognition. However, due to large variations in visual data, robust visual recognition is still an open problem. In recent years, sparse representation-based methods have become popular for visual recognition. By learning a compact dictionary of data and exploiting the notion of sparsity, start-of-the-art results have been obtained on many recognition tasks. However, existing data-driven sparse model techniques may not be optimal for some challenging recognition problems. In this dissertation, we consider some of these recognition tasks and present approaches based on sparse coding for robust and efficient recognition in such cases.
First we study the problem of low-resolution face recognition. This is a challenging problem, and methods have been proposed using super-resolution and machine learning based techniques. However, these methods cannot handle variations like illumination changes which can happen at low resolutions, and degrade the performance. We propose a generative approach for classifying low resolution faces, by exploiting 3D face models.
Further, we propose a joint sparse coding framework for robust classification at low resolutions. The effectiveness of the method is demonstrated on different face datasets.
In the second part, we study a robust feature-level fusion method for multimodal biometric recognition. Although score-level and decision-level fusion methods exist in biometric literature, feature-level fusion is challenging due to different output formats of biometric modalities. In this work, we propose a novel sparse representation-based method for multimodal fusion, and present experimental results for a large multimodal dataset. Robustness to noise and occlusion are demonstrated.
In the third part, we consider the problem of domain adaptation, where we want to learn effective classifiers for cases where the test images come from a different distribution than the training data. Typically, due to high cost of human annotation, very few labeled samples are available for images in the test domain. Specifically, we study the problem of adapting sparse dictionary-based classification methods for such cases. We describe a technique which jointly learns projections of data in the two domains, and a latent dictionary which can succinctly represent both domains in the projected low dimensional space. The proposed method is efficient and performs on par or better than many competing state-of-the-art methods.
Lastly, we study an emerging analysis framework of sparse coding for image classification. We show that the analysis sparse coding can give similar performance as the typical synthesis sparse coding methods, while being much faster at sparse encoding. In the end, we conclude the dissertation with discussions and possible future directions
Query optimization in mobile environments
No full textWe consider the issue of optimizing queries for distributed processing in mobile environment. An interesting characteristic of mobile machines is that they depend on battery as a source of energy which may not be substantial enough. Hence, the appropriate optimization criterion in a mobile environment considers both resource utilization and energy consum- ption at the mobile client.
In this scenario, the optimal plan for a query depends on the residual battery level of the mobile client and the load at the server. We approach this problem by compiling a query into a sequence of candidate plans, such that for any state of the client-server system, the optimal plan is one of the candidate plans.
A general solution is proposed by adapting the partial order dynamic programming search algorithm (p.o dp) such that the coverset of the query is the set of candidate plans. We propose two novel algorithms, namely, the linear combinations algorithm and the linearset algorithm (referred to as the linear algorithms) that compute the linearset of a query. The linear- set of a query is an approximation to the coverset returned by p.o. dp.
We show, by means of simulation, that (1) the linearset is an excellent approximation of the coverset, (2) query compilation using the linear algorithms outperform query compilation using p.o. dp by factors ranging from 2 to 9, (3) the time taken to compile queries using the linear algorithms for the general optimization criterion is at most twice the time taken by a System R* like standard query optimizer search algorithm, and (4) the run time overhead incurred by the linear algorithms technique is minimal.
The techniques presented in the paper are of general applicability to multi-criterion optimization problems in distributed databases, where each criterion is an additive metric.Technical report lcsr-tr-21
Interactive machine learning for complex graphics selection
No full textThesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2016.This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.Cataloged from student-submitted PDF version of thesis.Includes bibliographical references (pages 89-91).Modern vector graphics editors support the creation of a wonderful variety of complex designs and artwork. Users produce highly realistic illustrations, stylized representational art, even nuanced data visualizations. In light of these complex graphics, selections, representations of sets of objects that users want to manipulate, become more complex as well. Direct manipulation tools that artists and designers find accessible and useful for editing graphics such as logos and icons do not have the same applicability in these more complex cases. Given that selection is the first step for nearly all editing in graphics, it is important to enable artists and designers to express these complex selections. This thesis explores the use of interactive machine learning techniques to improve direct selection interfaces. To investigate this approach, I created Insight, an interactive machine learning selection tool for making a relevant class of complex selections: visually similar objects. To make a selection, users iteratively provide examples of selection objects by clicking on them in the graphic. Insight infers a selection from the examples at each step, allowing users to quickly understand results of actions and reactively shape the complex selection. The interaction resembles the direct manipulation interactions artists and designers have found accessible, while helping express complex selections by inferring many parameter changes from simple actions. I evaluated Insight in a user study of digital designers and artists, finding that Insight enabled users to effectively and easily make complex selections not supported by state-of-the-art vector graphics editors. My results contribute to existing work by both indicating a useful approach for providing complex representation access to artists and designers, and showing a new application for interactive machine learning.by Sumit Gogia.M. Eng
Optimizing queries for coarse grain parallelism
No full textWe consider the problem of optimizing select-project-join relational queries for minimum response time on parallel machines. The design of the optimizer is based on three ideas: (1) the concept and quantification of degree of coarse grain parallelism for an execution tree, (2) the design of a parallelizing scheduler for a tree of coarse grain operations which is provably near optimal, and (3) the analysis of the scheduling algorithm to obtain a cost formula for parallel execution time. The search algorithm of the optimizer is presented as a multi-dimensional dynamic programming algorithm. We present two three- dimensional search algorithms for the case when placement of relations in the parallel machine do not overlap. We propose the tree placement strategy and demonstrate, by means of examples, how the number of dimensions in the search can be significantly reduced, thereby increasing the efficiency of the search algorithm.Technical report lcsr-tr-21
Redox-Responsive Nanocapsules for the Spatiotemporal Release of Miltefosine in Lysosome: Protection against Leishmania
No full textLeishmaniasis, a vector-borne disease, is caused by intracellular parasite Leishmania donovani. Unlike most intracellular pathogens, Leishmania donovani are lodged in parasitophorous vacuoles and replicate within the phagolysosomes in macrophages. Effective vaccines against this disease are still under development, while the efficacy of the available drugs is being questioned owing to the toxicity for nonspecific distribution in human physiology and the reported drug-resistance developed by Leishmania donovani. Thus, a stimuli-responsive nanocarrier that allows specific localization and release of the drug in the lysosome has been highly sought after for addressing two crucial issues, lower drug toxicity and a higher drug efficacy. We report here a unique lysosome targeting polymeric nanocapsules, formed via inverse mini-emulsion technique, for stimuli-responsive release of the drug miltefosine in the lysosome of macrophage RAW 264.7 cell line. A benign polymeric backbone, with a disulfide bonding susceptible to an oxidative cleavage, is utilized for the organelle-specific release of miltefosine. Oxidative rupture of the disulfide bond is induced by intracellular glutathione (GSH) as an endogenous stimulus. Such a stimuli-responsive release of the drug miltefosine in the lysosome of macrophage RAW 264.7 cell line over a few hours helped in achieving an improved drug efficacy by 200 times as compared to pure miltefosine. Such a drug formulation could contribute to a new line of treatment for leishmaniasis.A. Das acknowledges SERB (India) Grants (CRG/2020/000492 and JCB/2017/000004) and DBT Grant (BT/PR22251/NNT/28/1274/2017) for supporting this research. N. Mukherjee acknowledges SERB (India) Grant PDF/2016/001437 and K. Das acknowledges the grant EMR/2015/001674 for supporting this research. Financial support from DST (DST/INSPIRE/03/2017/002477) is acknowledged by R.T. This manuscript bears CSMCRI registration no 7/2021.Pramanik, SK (corresponding author), CSIR Cent Salt & Marine Chem Res Inst, Bhavnagar 364002, Gujarat, India.
Mukherjee, N (corresponding author), CSIR Indian Inst Chem Biol, Canc Biol & Inflammatory Disorder Div, Kolkata 700032, India.
Chattopadhy, S (corresponding author), BITS Pilani, Pilani 403726, Goa, India.
Das, A (corresponding author), Indian Inst Sci Educ & Res Kolkata, Mohanpur 741246, W Bengal, India.
[email protected]; [email protected]; [email protected]
Rola innowacji w zrównoważonym systemie sanitarnym: studium przypadku Indii
No full textSanitation and water are one of those problems which have been given top priority in the sustainable agenda. However, scanty resources, geographical condition, natural environment, tradition, institutional and financial constraints lead to several challenges of feasibility, affordability, availability,and acceptability. This study reveals the inequality in the access to improved toilet facilities based on wealth index and locality of households using National Family Health Survey (NFHS) data. These problems can be addressed by applying different types of social innovations in which novelty in product and process can play a crucial role. This paper critically examines the role of innovation which can play in expanding transition to sustainable development in the sanitation sector which needs some financial, organizational, and institutional agreement. The progress in sanitation sector is dependent on the consumer behavior. However, it still lacks a variety of quality-price ranges and its utility as the basic needs of dignified life.Warunki sanitarne i woda to jedne z najważniejszych wyzwań w kontekście zrównoważonego rozwoju. Zarazem skąpe zasoby, warunki geograficzne, środowisko naturalne, tradycja, ograniczenia instytucjonalne i finansowe prowadzą do kilku wyzwań związanych z wykonalnością, przystępnością cenową, dostępnością i akceptowalnością. Badanie to ujawnia nierówności w dostępie do ulepszonych toalet w oparciu o indeks zamożności i lokalizację gospodarstw domowych na podstawie danych National Family Health Survey (NFHS). Problemy te można rozwiązać, stosując różne rodzaje innowacji społecznych, w których nowość w produkcie i procesie może odgrywać kluczową rolę. W artykule krytycznie przeanalizowano rolę innowacji, które mogą odegrać istotną rolę w przejściu do zrównoważonego rozwoju w sektorze sanitarnym, które wymaga finansowego, organizacyjnego i instytucjonalnego zabezpieczenia. Postęp w sektorze sanitarnym zależy też od zachowań konsumentów. Jednak nadal brakuje tu różnych przedziałów jakościowo-cenowych i użyteczności zapewniających podstawowe potrzeby godnego życia
Gendered Disparities in Water and Sanitation through an Intersectional Lens: Emphasising Women’s Perspectives
No full textThe relationship between gender and water, sanitation, and hygiene (WASH) has been a widely accepted concern among academicians, activists, and social workers in India, but research and policies focusing on gender and sanitation often fail to address the issue of the intersectionality of identities. Analysing the complex intersection of caste, class, age and gender with water and sanitation in rural India extends new opportunities to explore the complex power dynamics prevalent in society. A focus group study with 54 female participants of seven discussions and in-depth interviews has been conducted in the Hardoi district of Uttar Pradesh to explore the social relations and differences in the physical world within the context of accessibility, affordability, and availability in the water and sanitation sector. Given gendered and other social divisions, we elaborate on how women play an essential role in water and sanitation management in the household. This study also offers evidence of rural women’s experiences of intra-personal, household, and social harassment and violence related to poor or absence of sanitation and water infrastructure due to intersectional social dynamics. We also demonstrate how theorising about a single dimension of social difference ignores the different groups’ access to power, leading to inequality and disparity.
Image and video processing based on intrinsic attributes
No full textAdvancements in computer vision techniques driven by machine learning have facilitated robust and efficient estimation of attributes such as depth, optical flow, albedo, and shading. To encapsulate all such underlying properties associated with images and videos, we evolve the concept of intrinsic images towards intrinsic attributes. Further, rapid hardware growth in the form of high-quality smartphone cameras, readily available depth sensors, mobile GPUs, or dedicated neural processing units have made image and video processing pervasive. In this thesis, we explore the synergies between the above two advancements and propose novel image and video processing techniques and systems based on them. To begin with, we investigate intrinsic image decomposition approaches and analyze how they can be implemented on mobile devices. We propose an approach that considers not only diffuse reflection but also specular reflection; it allows us to decompose an image into specularity, albedo, and shading on a resource constrained system (e.g., smartphones or tablets) using the depth data provided by the built-in depth sensors. In addition, we explore how on-device depth data can further be used to add an immersive dimension to 2D photos, e.g., showcasing parallax effects via 3D photography. In this regard, we develop a novel system for interactive 3D photo generation and stylization on mobile devices. Further, we investigate how adaptive manipulation of baseline-albedo (i.e., chromaticity) can be used for efficient visual enhancement under low-lighting conditions. The proposed technique allows for interactive editing of enhancement settings while achieving improved quality and performance. We analyze the inherent optical flow and temporal noise as intrinsic properties of a video. We further propose two new techniques for applying the above intrinsic attributes for the purpose of consistent video filtering. To this end, we investigate how to remove temporal inconsistencies perceived as flickering artifacts. One of the techniques does not require costly optical flow estimation, while both provide interactive consistency control. Using intrinsic attributes for image and video processing enables new solutions for mobile devices – a pervasive visual computing device – and will facilitate novel applications for Augmented Reality (AR), 3D photography, and video stylization. The proposed low-light enhancement techniques can also improve the accuracy of high-level computer vision tasks (e.g., face detection) under low-light conditions. Finally, our approach for consistent video filtering can extend a wide range of image-based processing for videos.Fortschritte im Bereich der Computer-Vision-Techniken, die durch Maschinelles Lernen vorangetrieben werden, haben eine robuste und effiziente Schätzung von Attributen wie Tiefe, optischer Fluss, Albedo, und Schattierung ermöglicht. Um all diese zugrundeliegenden Eigenschaften von Bildern und Videos zu erfassen, entwickeln wir das Konzept der intrinsischen Bilder zu intrinsischen Attributen weiter. Darüber hinaus hat die rasante Entwicklung der Hardware in Form von hochwertigen Smartphone-Kameras, leicht verfügbaren Tiefensensoren, mobilen GPUs, oder speziellen neuronalen Verarbeitungseinheiten die Bild- und Videoverarbeitung allgegenwärtig gemacht. In dieser Arbeit erforschen wir die Synergien zwischen den beiden oben genannten Fortschritten und schlagen neue Bild- und Videoverarbeitungstechniken und -systeme vor, die auf ihnen basieren. Zunächst untersuchen wir intrinsische Bildzerlegungsansätze und analysieren, wie sie auf mobilen Geräten implementiert werden können. Wir schlagen einen Ansatz vor, der nicht nur die diffuse Reflexion, sondern auch die spiegelnde Reflexion berücksichtigt; er ermöglicht es uns, ein Bild auf einem ressourcenbeschränkten System (z. B. Smartphones oder Tablets) unter Verwendung der von den eingebauten Tiefensensoren bereitgestellten Tiefendaten in Spiegelung, Albedo und Schattierung zu zerlegen. Darüber hinaus erforschen wir, wie geräteinterne Tiefendaten genutzt werden können, um 2D-Fotos eine immersive Dimension hinzuzufügen, z. B. um Parallaxen-Effekte durch 3D-Fotografie darzustellen. In diesem Zusammenhang entwickeln wir ein neuartiges System zur interaktiven 3D-Fotoerstellung und -Stylisierung auf mobilen Geräten. Darüber hinaus untersuchen wir, wie eine adaptive Manipulation der Grundlinie-Albedo (d.h. der Farbintensität) für eine effiziente visuelle Verbesserung bei schlechten Lichtverhältnissen genutzt werden kann. Die vorgeschlagene Technik ermöglicht die interaktive Bearbeitung von Verbesserungseinstellungen bei verbesserter Qualität und Leistung. Wir analysieren den inhärenten optischen Fluss und die zeitliche Konsistenz als intrinsische Eigenschaften eines Videos. Darüber hinaus schlagen wir zwei neue Techniken zur Anwendung der oben genannten intrinsischen Attribute zum Zweck der konsistenten Videofilterung vor. Zu diesem Zweck untersuchen wir, wie zeitliche Inkonsistenzen, die als Flackerartefakte wahrgenommen werden, entfernt werden können. Eine der Techniken erfordert keine kostspielige optische Flussschätzung, während beide eine interaktive Konsistenzkontrolle bieten. Die Verwendung intrinsischer Attribute für die Bild- und Videoverarbeitung ermöglicht neue Lösungen für mobile Geräte - ein visuelles Computergerät, das aufgrund seiner weltweiten Verbreitung von großer Bedeutung ist - und wird neuartige Anwendungen für Augmented Reality (AR), 3D-Fotografie und Videostylisierung ermöglichen. Die vorgeschlagenen Low-Light-Enhancement-Techniken können auch die Genauigkeit von High-Level-Computer-Vision-Aufgaben (z. B. Objekt-Tracking) unter schlechten Lichtverhältnissen verbessern. Schließlich kann unser Ansatz zur konsistenten Videofilterung eine breite Palette von bildbasierten Verarbeitungen für Videos erweitern
Experience Individualization on Online TV Platforms through Persona-based Account Decomposition
No full text- …
