Current research focus is on (a) integration of human and contextual information in analyzing images and video, leading to bio-inspired methods for computer vision; (b) large scale camera networks and associated "big-data" information processing tasks; and (c) bio-medical image informatics and brain connectomics. In addition, we continue the work on fundamental problems of segmentation, registration and tracking, and large scale image/video indexing and search, with emphasis on developing robust and scalable methods.
The BisQue image informatics infrastructure, developed by the VRL group and the Center for Bio-Image Informatics/Center for Multimodal Bigdata Science and Healthcare is now available as a core service in the CyVerse cyberinfrastructure. BisQue is currently being used in many major laboratories around the world and we are in the process of adding image feature services and indexing services to facilitate large scale image data management. The new release will also include deep learning into the image feature extraction and recognition workflow.
Distinguished Professor and Chair
Department of Electrical and Computer Engineering
University of California
Santa Barbara, CA 93106-9560
Tel: (805) 893 7112
E-mail: manj [at] ucsb [dot] edu
Computer Vision and Machine Learning
Over the years VRL researchers have contributed to the development of robust methiods for image/video segmentation, classification, and activity recognition. These methods have been applied to various application scenarios, including aerial/satellite images, activity detection in a hospital ICU and detecting complex activities in distributed camera networks.
Security and Media Forensics
Research in the lab has focused on 2 different problems: (1) one relating to computer malware where we introduced novel signal/image processing methods for detecting and classifying malware at accuracies comparable to or exceeding state of the art traditional methods, and orders of magnitude faster. (2) VRL is one of the early groups to develop methods for detecting and localizing image manipulations. The methods developed by the lab members, together with industry and academic partners, have consistently been ranked among the top performing methods in large scale test and evaluation for image manipulation detection.
BisQue Software Infrastructure
Managing large amount of multimodal data is a significant problem in many scientific applications. The software infrastructure BisQue developed at UCSB is a unique open source resource that combines data management with image analysis and machine learning that enables reproducible computer vision. BisQue can be deployed easily in a cloud computing environment and developers can add customized computational modules and workflows, all within a standard web-browser based interface, to carry out computations at scale. In addition to core image management services (data upload, visualization, annotation, sharing and analysis), current research focuses on application to materials science, marine science and health/medical imaging data.