Abstract: In this tutorial, I will give an overview of the progress towards the goal of single-view 3D inference i.e. being able to infer the 3D structure underlying a single input image. I will briefly summarize the different approaches proposed over the past decades and then focus on more recent learning based methods that tackle this task. Towards building more scalable approaches, we will categorize these learning based methods in context of the supervision required, and I will show how incorporating geometry in the learning process can relax the need of strong supervision. I will finally highlight some exciting future directions and open research questions.

Bio: Shubham Tulsiani is a research scientist at Facebook AI Research (FAIR) and will be joining the CMU School of Computer Science as an Assistant Professor in Fall 2021. He received a PhD. in Computer Science from UC Berkeley under the supervision of Jitendra Malik in 2018. He is interested in building perception systems that can infer the spatial and physical structure of the world they observe. His work was awarded the Best Student Paper Award at CVPR 2015.

Title of the talk: Computational Neuroimaging in Pathology: Basics and Scope

Date&Time: 20 December, 11:00 AM (IST).
Session Chair: Arnav Bhavsar

With neurological and neuro-psychiatric disorders on the rise each year, there has been an increased impetus to understand and quantify the complex information conveyed through brain magnetic resonance images (MRI) for reliable prognosis, diagnosis and for developing personalized treatment strategies for better efficacy and outcomes. Quantitative analysis of brain MRI can unravel the brain structure, function and architecture that can facilitate deeper insights into the neurobiological underpinnings of the pathology under consideration. My talk will focus on computational neuroimaging that includes pre-processing and analysis of these images that is crucial for precise quantification to overcome the limitations of subjective visual interpretation, consequently improving the workflow and reliability in radiology and supporting neurology based decision making.

Bio: Dr. Madhura Ingalhalikar, has cutting-edge experience in the field of medical image analysis for the past 13 years. She is well versed with MRI protocols, expert in processing pipelines, multi-modal MRI analysis, machine learning and deep learning. Her work on Diffusion MRI at UPenn produced multiple high impact journal papers including a paper in PNAS which was commended by the neuro-imaging community and received great media attention. Currently, she is a Professor and heads the Symbiosis Centre for Medical Image analysis (SCMIA), Pune, and focuses on using multi-modality imaging and advanced state of the art deep learning techniques in prediction and classification problems of several neurological disorders with a special emphasis on brain tumors and movement disorders.

Title of the talk: Entering the statistical domain: Use of deformable statistical shape modeling in clinical practices.

Date&Time: 20 December, 12:00 Noon (IST).
Session Chair: Arnav Bhavsar

Abstract: Medicine is evolving rapidly with emerging applications leveraging the fields of machine learning, statistics, image analysis, and biomechanics. Statistical inferences in terms of statistical shape modeling have allowed distributed diagnosis on the organs' anatomical shape (geometry), biological and population variability, and its representation in imaging modalities. Variation in size is also an important determinant for variation in many other organismal traits. The interface of statistical shape modeling of the human joints has been recently recognized within the research community. Once developed, Statistical Shape Models (SSMs) and Statistical Appearance Models (SAMs) have various applications in biomechanics including automatic image segmentation, subject-specific biomechanics models, population classifications based on bone morphometry, relationship between form or shape and function etc. The techniques involved are emerging and have promising applications towards modeling patient-specificity. This tutorial would focus on explaining the science and mathematics that goes into building statistical shape models for solving clinical problems, validations conducted on such methods, or applications of such methods to solve specific clinical problems.

Bio: Dr. Bhushan Borotikar has been a translational researcher in the field of medical sciences since 2005. His research spans the domains of medicalimaging, computational anatomy, human movement sciences, machine learning, and orthopedic biomechanics. Dr. Bhushan Borotikar is currently engaged as a research faculty(Associate Professor) at the Symbiosis Centre for Medical Image Analysis,Symbiosis International University, Pune, India. He also has adjunct appointments at the Laboratory for the Treatment of Medical Information (LaTIM),INSERM, U1101, Brest, France as a Senior Research Scientist and at the Division of Biomedical Engineering, University of Cape Town, Cape Town, South Africa, asan honorary Associate Professor. Dr Borotikar finished his PhD from Cleveland State University, Ohio, USA and Cleveland Clinic, Ohio, USA in 2009.He then worked as a post-doctoral fellow at the National Institutes of Health(NIH), Bethesda, USA till 2014. From 2014 till 2020, he was at the Frenchlaboratory of excellence- LaTIM. Dr Borotikar's current research interests are focused on developing clinical tools and procedures using emerging engineering technologies in three domains: 1) Musculoskeletal disorders and sports injuries in adults and children, 2) Computer assisted orthopaedic surgeries, joint implants and medical device designs, and 3) Advances in medical image acquisition,analysis, and synthesis. Dr Borotikar has published extensively in peer-reviewed journals, international conferences, has written book chapter(s) on his research, and has a provisional patent application on his name.

Title: Marriage of Computer Vision, Speech and Natural Language

Date&Time: 19 December, 11:00 AM (IST).
Session Chair: Aditya Nigam

Abstract: Speech generally is considered to have three parts to it: vision, aural, and the social construct. In recent years, although the field has been moving at a dramatic pace, progress is being made in silos. The primary reason for this being that speech is considered "spoken text" by practitioners and researchers alike. Most open-source datasets due to their distance from real-world conditions help in spreading this false impression. In this condition, it is not surprising that common and important features of speech like intonation and disfluency do not get captured by this intent. This tutorial aims to provide an appreciation of the "full-stack" of speech - aural, vision and the textual (or social construct) parts with a special emphasis on aspects that may have significance for current and future research.

Bio: Rajiv Ratn Shah currently works as an Assistant Professor in the Department of Computer Science and Engineering (joint appointment with the Department of Human-centered Design) at IIIT-Delhi. He is also the director of MIDAS lab at IIIT Delhi. He received his Ph.D. in computer science from the National University of Singapore, Singapore. Before joining IIIT-Delhi, he worked as a Research Fellow in Living Analytics Research Center (LARC) at the Singapore Management University, Singapore. Prior to completing his Ph.D., he received his M.Tech. and M.C.A. degrees in Computer Applications from the Delhi Technological University, Delhi and Jawaharlal Nehru University, Delhi, respectively. He has also received his B.Sc. in Mathematics (Honors) from the Banaras Hindu University, Varanasi. Dr. Shah is the recipient of several awards, including the prestigious Heidelberg Laureate Forum (HLF) and European Research Consortium for Informatics and Mathematics (ERCIM) fellowships. He is involved in organizing and reviewing of many top-tier international conferences and journals. His research interests include multimedia content processing, natural language processing, image processing, speech processing, multimodal computing, data science, and social media computing.

Yaman Kumar is a PhD student at IIIT Delhi and University of Buffalo, USA. Recently, he has received a Google PhD fellowship. Prior to joining his PhD program, Yaman had done his BTech from NSIT, Delhi and was working with Adobe systems. He is also the recipient of several awards at top conferences such as the best student paper award at AAAI 2019.