1,501 Hits in 9.4 sec

This Looks Like That... Does it? Shortcomings of Latent Space Prototype Interpretability in Deep Networks [article]

Adrian Hoffmann, Claudio Fanconi, Rahul Rade, Jonas Kohler
2021 arXiv   pre-print
Among these networks, the arguably most widespread approach is so-called prototype learning, where similarities to learned latent prototypes serve as the basis of classifying an unseen data point.  ...  Specifically, we find that this network's interpretability mechanism can be led astray by intentionally crafted or even JPEG compression artefacts, which can produce incomprehensible decisions.  ...  JPEG compression is a lossy compression algorithm, which leaves artefacts behind.  ... 
arXiv:2105.02968v4 fatcat:av2g4y2k7fb4tjtfbksdiujasy

Deep Learning-Based Video Coding: A Review and A Case Study [article]

Dong Liu, Yue Li, Jianping Lin, Houqiang Li, Feng Wu
2019 arXiv   pre-print
In the hope of advocating the research of deep learning-based video coding, we present a case study of our developed prototype video codec, namely Deep Learning Video Coding (DLVC).  ...  Both tools help improve the compression efficiency by a significant margin.  ...  ACKNOWLEDGMENT This work is supported by the Natural Science Foundation of China under Grant 61772483.  ... 
arXiv:1904.12462v1 fatcat:cgcmwylwnvd67abpl6wders5hy

A two-domain real-time algorithm for optimal data reduction: a case study on accelerator magnet measurements

Pasquale Arpaia, Marco Buzio, Vitaliano Inglese
2010 Measurement science and technology  
The second technique uses a data reduction algorithm in order to improve the compression ratio while preserving the consistency of the measured signal.  ...  A real-time algorithm of data reduction, based on the combination a two lossy techniques specifically optimized for high-rate magnetic measurements in two domains (e.g. time and space), is proposed.  ...  Acknowledgments This work is supported by CERN trough the agreement LHC/AT/K1464 with the University of Sannio, whose supports authors gratefully acknowledge.  ... 
doi:10.1088/0957-0233/21/3/035103 fatcat:imvchsxgabavvlrb5jkjwtng6a

Fast downscaled inverses for images compressed with M-channel lapped transforms

R.L. de Queiroz, R. Eschbach
1997 IEEE Transactions on Image Processing  
In this fashion, fewer transform coefficients are used and the synthesis process is simplified. We studied the design of fast inverses, for a given forward transform.  ...  We studied downscaled inverses where the image is decompressed partially, and a reduced inverse transform is used to recover the image.  ...  CONCLUSION We studied methods to decompress a compressed image to a lower resolution.  ... 
doi:10.1109/83.585231 pmid:18282974 fatcat:5shxlec5wrh5tjlsyp7a2ifdma

Prototyping pipelined applications on a heterogeneous FPGA multiprocessor virtual platform

Antonino Tumeo, Marco Branca, Lorenzo Camerini, Marco Ceriani, Matteo Monchiero, Gianluca Palermo, Fabrizio Ferrandi, Donatella Sciuto
2009 2009 Asia and South Pacific Design Automation Conference  
We discuss several case studies, analyzing the interaction of the architecture and applications and we show that multimedia and telecommunication applications with unbalanced pipeline stages can be easily  ...  We introduce a prototyping platform (FPGAbased) and a methodology for these applications. Our platform consists of a mix of standard and custom heterogeneous cores.  ...  ACKNOWLEDGMENTS Research partially funded by the European Community's Sixth Framework Programme, hArtes Project.  ... 
doi:10.1109/aspdac.2009.4796500 dblp:conf/aspdac/TumeoBCCMPFS09 fatcat:5sth3x5itfacti7zpnj3qhup3y

Combining on-hardware prototyping and high-level simulation for DSE of multi-ASIP systems

Paolo Meloni, Sebastiano Pomata, Luigi Raffo, Roberta Piscitelli, Andy D. Pimentel
2012 2012 International Conference on Embedded Computer Systems (SAMOS)  
The approach is then evaluated with a design space exploration case study, taking as a target application a video compression kernel.  ...  As a possible solution, within the MADNESS project [1], an integrated toolset has been proposed, combining the benefits of novel fast FPGA-based prototyping techniques with those provided by high-level  ...  The target application is a motion-JPEG (M-JPEG in the following) video compression kernel.  ... 
doi:10.1109/samos.2012.6404191 dblp:conf/samos/MeloniPRPP12 fatcat:rzmvcr6x25hjnf42bozaajw5fm

Development and evolution of a heterogeneous continuous media server: a case study

Dwight J. Makaroff, Norman C. Hutchinson
2005 Journal of Software Maintenance and Evolution Research and Practice  
This paper provides a case study of the design, implementation, and evolution of a continuous media file server.  ...  Since continuous media servers must send the raw data to a client application over a network, the protocol considerations, hardware interface and data storage/retrieval methods are of the paramount importance  ...  ACKNOWLEDGEMENTS The authors would like to thank the members of the CMFS development design and implementation team, most notably the Principal Investigator, Dr.  ... 
doi:10.1002/smr.309 fatcat:stm5wat5szfgxnn52qdjabiehm

How real is real enough? optimal reality sampling for fast recognition of mobile imagery

Antti Oulasvirta, Antti Nurminen, Tiia Suomalainen
2012 ACM Transactions on Applied Perception  
Resolution, image size, and JPEG compression of images of façades were manipulated in a same-di↵erent recognition task carried out in the field.  ...  In particular, we identify the minimum information required for fast recognition of images of directly perceivable real-world buildings displayed on a mobile device.  ...  If the data must remain inviolate, a lossless compression algorithm may be employed. However, if the service is targeted at human receivers, a lossy algorithm may be applicable.  ... 
doi:10.1145/2355598.2355604 fatcat:lhynfzwfcfbp7pbzm25d5vnk5i

Energy-Aware Load Direction for Servers: A Feasibility Study

Shane Case, Furat Afram, Erdem Aktas, Kanad Ghose
2012 2012 20th Euromicro International Conference on Parallel, Distributed and Network-based Processing  
The technique requires no a priori knowledge about individual workloads and manages the server states explicitly.  ...  We use synthetic scripts, a small server setup and actual energy consumption measurements to show that the proposed system achieves non-trivial energy savings at typical operating regions of traditional  ...  Acknowledgements This work is supported in part by NYSERDA under agreement number 10946 and by the NSF awards, numbers 0958501 and 1040666.  ... 
doi:10.1109/pdp.2012.63 dblp:conf/pdp/CaseAAG12 fatcat:7xswsbuwwjewbbdmduamkvoaha

System-Level Performance Evaluation of Very High Complexity Media Applications : A H264/AVC Encoder Case Study

Hajer Krichene Zrida, Abderrazek Jemai, Ahmed C Ammari, Mohamed Abid
2011 International Journal of Communications, Network and System Sciences  
Raising the abstraction level is nowadays widely seen as a solution to bridge the gap between the increasing system complexity and the low design productivity.  ...  For this purpose, we have selected a system level design of a very high complexity media application; a H.264/AVC (Advanced Video Codec) video encoder.  ...  Medium Complexity Media Case Studies The Sesame modeling and simulation methodology has been applied to two medium complexity media applications: an MPEG-2 decoder and a variant of an M-JPEG encoder [  ... 
doi:10.4236/ijcns.2011.47052 fatcat:egxzq3wz7vhp3f5ggadstywoxa

Learning Fast Algorithms for Linear Transforms Using Butterfly Factorizations [article]

Tri Dao, Albert Gu, Matthew Eichhorn, Atri Rudra, Christopher Ré
2020 arXiv   pre-print
On a standard task of compressing a single hidden-layer network, our method exceeds the classification accuracy of unconstrained matrices on CIFAR-10 by 3.9 points – the first time a structured approach  ...  We ask to what extent hand-crafting these algorithms and implementations is necessary, what structural priors they encode, and how much knowledge is required to automatically learn a fast algorithm for  ...  , of DARPA, NIH, ONR, or the U.S.  ... 
arXiv:1903.05895v2 fatcat:v7iaichnozgnvbnrn4vqvka5le

Fast Wavelet Image Denoising Based On Local Variance And Edge Analysis

Gaoyong Luo
2008 Zenodo  
The proposed algorithm provides a potential solution with parallel computation for real time or embedded system application.  ...  Fast lifting wavelet spatial-frequency decomposition and reconstruction is developed with the advantages of being computationally efficient and boundary effects minimized.  ...  The algorithm developed uses weighted coefficients of lifting factorization of 1D prototype transform, but replacing 1D neighborhoods by 2D rings [27] .  ... 
doi:10.5281/zenodo.1081654 fatcat:tlj5y2yg4re3foxuoxb64ua24y

A Rapid Prototyping Environment for Wireless Communication Embedded Systems

Bryan A. Jones, Joseph R. Cavallaro
2003 EURASIP Journal on Advances in Signal Processing  
The papers in this special issue span a broad range of topics related to the rapid prototyping of DSP systems.  ...  The primary goal of the emulator is to support design space exploration of real-time algorithms.  ...  ACKNOWLEDGMENTS ACKNOWLEDGMENT We would like to thank Aptix Corporation (San Jose, Calif, USA) for the technical support during the prototype implementation of our methodology using the Aptix system  ... 
doi:10.1155/s111086570330304x fatcat:3ird7hyyzjeb5bmrqa6ztx7574

Pi of the Sky – all-sky, real-time search for fast optical transients

A. Burd, M. Cwiok, H. Czyrkowski, R. Dabrowski, W. Dominik, M. Grajda, M. Husejko, M. Jegier, A. Kalicki, G. Kasprowicz, K. Kierzkowski, K. Krupska (+12 others)
2005 New Astronomy  
Performance of the prototype is briefly reviewed and perspectives for the future are outlined.  ...  The prototype with two cameras has been installed at Las Campanas (Chile) and is operational from July 2004. The paper presents general idea and describes the apparatus in detail.  ...  This project from the very beginning benefits from the experience gained with All Sky Automatic Survey (ASAS), generously shared with us by its leader, G. Pojmanski.  ... 
doi:10.1016/j.newast.2005.02.002 fatcat:vpk23k3wmba4lnlndvtkqs2kda

Application-Aware Acceleration for Wireless Data Networks: Design Elements and Prototype Implementation

Zhenyun Zhuang, Tae-Young Chang, R. Sivakumar, A. Velayutham
2009 IEEE Transactions on Mobile Computing  
We also build a proof-of-concept prototype of A3 which can work with real applications. ò Voice Over IP The expected VoIP call capacity in a one-hop IEEE 802.11b network with G.711 voice codec is about  ...  Using insights gained through the aforementioned study, we present a set of new transport principles in the form of deployable solution suite called wireless P2P (wP2P) that addresses the issues using  ...  However, it is a better way of identifying data-sets that need to be operated on by a given compression algorithm. IV. SOLUTION A.  ... 
doi:10.1109/tmc.2009.52 fatcat:wmvwzlqpsjbxvnccepn6euaa64
« Previous Showing results 1 — 15 out of 1,501 results