Challenges in Image Informatics Segmentation to SteganographyOutlineImage InformaticsPowerPoint PresentationSlide 5VRL & Image Informaticshttp://vision.ece.ucsb.eduCurrent & Recent SponsorsExisting centers and programsInteractive Digital MultimediaCenter for Bio-Image InformaticsCenter for Bioimage InformaticsOutline: current researchImage & Video RegistrationAerial ImagesVideo Mosaickinganother exampleOne more..(retina)ApplicationsImage DatabasesSlide 21Similarity search in images/videoQuery by ExampleSearch by Texture LayoutBeyond Similarity SearchBiomolecular Image DatabasesExample Data CollectionsMammalian retina is a complex structureRetinal Images: Curent workImage Databases: researchISO/MPEG-7 standardObjectives of MPEG-7Application ExamplesScope of MPEG-7UCSB descriptors in MPEG7Image SegmentationSlide 37Example – Medical ImageExample – Blood CellsSlide 40Example – Natural ImageSegmentation: natural imagesGeneral FrameworkGeneral Framework – Vector FieldBoundary ConcavitiesEdgeflow-based Diffusion ExamplesHiding Data in Images SteganographyHiding text in imageSlide 49Multimedia Data Hiding: A multidisciplinary fieldA data hiding systemDesign IssuesOther Design IssuesTrade-offsData Hiding at VRLHigh Volume Data HidingMethodologyWhy Image-adaptation is needed?Zero-threshold SEC scheme exampleZero-threshold SEC scheme example (II)Unity-threshold SEC scheme exampleImage Tampering ExamplesSlide 63Print-Scan Resilient Data HidingSelective Embedding in Low Frequencies (SELF)Undoing RotationZoomed image and its DFTRotated imageExample 1: Baboon imageExample 1 (cont.)Slide 71Example 2: Man imageSlide 73Example 2 (cont.)Robustness against other attacksSlide 76SteganalysisSummarySlide 79Retinal detachmentVertebrate retinaChallenges in Image InformaticsSegmentation to SteganographyB. S. ManjunathElectrical and Computer Engineering DepartmentUniversity of California, Santa [email protected]://vision.ece.ucsb.eduDr. Charles KenneyDr. Baris SumengenDmitry FederovKaushal SolankiJiyun ByunMotaz El SabanSitaram BhagavathyKen SullivanIEEENov 23, 20042Outline•Image Informatics: challenges ahead•Center for image informatics at UCSB?•Partnership with local industry leaders..•Vision Research Lab @ ECE, UCSB•Image Processing/Computer Vision Research at VRL•Image databases & standards•Image Segmentation•SteganographyIEEENov 23, 20043Image Informatics•images/video are ubiquitous•Many diverse applications•Biology, medical imaging, remote sensing, manufacturing, security/surveillance,..•common informatics problems in processing, pattern recognition, and databases•We plan to create a central home at UCSB for image informatics research.•exploit the synergy among science applications, to develop robust tools and technologies.IEEENov 23, 20044Biology, Medicine• High-speed nanoimaging• Cell classification• Tracking dynamic structures• Understanding cell birth and death• fMRI brain image analysis• Content-based access• Quantitative descriptorsMarine Science, Oceanography• Biological sampling• Species and their habitat classification• Study of pathogen transport • Underwater imaging instrumentationRemote sensing, Geo-surveillance• Automated analysis • Detecting patterns (e.g., airfields, harbors)• Semantic queries• Information integrationSecurity, Surveillance• Detecting illicit activity • Reconnaissance • Person tracking • Activity analysis• Face recognitionPeople• K-12 • Teachers • Undergraduates• Graduate students• Postdocs• Industries• FacultiesTechnology Transfer• New industry opportunities• Image library infrastructure• Adaptive image processing methods• Nanoimaging instrumentation • Underwater imaging instrumentation• Imaging sensor networksEducation& Outreach• CSU-SB• CSUN• Ventura College• New Media Studio Manufacturing• High-speed nanoimaging• Defect detection• Object recognitionCI3 ERCResearch Thrusts1. Imaging & Image Processing 2. Imaging Networks3. Distributed Learning, Pattern Recognition, Data Mining4. Image LibrariesTestbedsA. High-speed Plasmonic Nanoscale Optical Microscope (PNOM)B. Imaging Networks: OceaNET,SurveillanceNETC. Large-Scale Image Libraries• Why CI3? Images everywhere, diverse applications, common issues.• Goal: To radically advance image informatics in a synergetic manner.• How? An integrated approach to imaging, information processing, & applications, leading to new theories and methods.The diversity and ubiquitous nature of images in the critical path of many science applications makes this CI3 ERC a very timely and urgently needed effortERC: Center for Intelligent Image Informatics--Vision & Goals Participants: University of California at Santa Barbara, UC Berkeley, Woods Hole Oceanographic InstituteManjunath/UCSBIEEENov 23, 20045Knowledge BaseTechnology BaseTechnology IntegrationRT3: Learning, Pattern Recognition/Data MiningDistributed PR, Mining, Clustering & ClassificationRT1: Imaging/Image ProcessingNano-Imaging, Features,Segmentation,Optical Plankton SensorRT2: Imaging NetworksCamera networks,Sensor networksTestbed ANanoscale Optical Imaging: High Speed Plasmonic Nanoscale Optical MicroscopeTestbed BImaging Networks Underwater: OceaNET Security/Surveillance: SurveillanceNETTestbed CLarge ScaleImage Library Remote sensing BiologyRequirements Products & OutcomesRT4: Image Libraries/DatabasesIntegration, Scalable access, Incorporating uncertainty,Domain semantics CI3 Strategic Planning ChartManufacturing Tracking Object recognition  Object quantification Search and retrievalOceanographyRemoteSensingHomelandSecurityManjunath/UCSBCore research technologiesApplication enablingtechnologiesApplicationdemonstrationsGraduate studentsPostdocs, Summer internsResearch engineersIEEENov 23, 20046VRL & Image Informatics•Leading research in image databases already happening•Alexandria Digital Library (NSF, 1994-2004)•Center for Bio-Image Informatics (NSF, 2003-08)•MPEG-7 work (Samsung, 1998-2001)•IGERT in Interactive Digital Multimedia (NSF, 2003-08)IEEENov 23, 20047•Image Texture Analysis•Image and video segmentation•Image/Video Registration•Image/Video Databases•content based retrieval•data mining•High dimensional indexing and search•Data hiding/Steganographyhttp://vision.ece.ucsb.eduIEEENov 23, 20048•Office of Naval Research•Marine Corps•Chinalake Navy Lab•National Science Foundation