About

B.S. Manjunath is a Distinguished Professor and the Frank Koenig Chair in the Department of Electrical and Computer Engineering at UC Santa Barbara. His research interests include image informatics, with a focus on the integration of imaging and data analysis to advance scientific and healthcare applications. He is associated with the Vision Research Lab and is involved with centers such as the Center for Bio-Image Informatics and the Center for Multimodal Big Data Science and Healthcare. His work emphasizes the development of innovative methods for image analysis and informatics, contributing to the fields of bio-image informatics and big data in healthcare.

Research topics

• Computer Science
• Artificial Intelligence
• Machine Learning
• Computer vision
• Data Mining
• Theoretical computer science
• Speech recognition

Selected publications

• EVALUATION OF MODIFIED SHOCK INDEX AS AN INDICATOR FOR PROGNOSIS AMONG SEPSIS PATIENTS WITH AND WITHOUT COMORBIDITIES PRESENTING TO EMERGENCY DEPARTMENT: A PROSPECTIVE OBSERVATIONAL STUDY

  International Journal of Advanced Research · 2025-07-31

  articleOpen access

  Background and Aim: Despite the established association between sepsis an pre-existing comorbid conditions, limited information is available about the survival impact on sepsis patients.We aimed to assess the predictive ability of modified shock index (MSI) among sepsis patients with and without comorbidities presenting to the emergency department (ED) at our tertiary care centre. Methodology: This is a prospective, observational study consisting of total 140 patients presented with sepsis. MSI was calculated by dividing the heart rate(HR) by the mean arterial pressure(MAP),and determined at the time of admission. The predictive ability, sensitivity, and specificity of MSI was evaluated. Conclusion: MSI had a good predictive ability for sepsis patients without comorbid conditions. While MSI carries suboptimal predictive ability for sepsis patients with comorbid conditions. Owing to MSI, a simple index that can be calculated at the bedside, can be utilized in the emergency department for the management of sepsis patients without comorbidities.

  PublisherDOI

• RareSpot: Spotting Small and Rare Wildlife in Aerial Imagery with Multi-Scale Consistency and Context-Aware Augmentation

  ArXiv.org · 2025-06-23 · 1 citations

  preprintOpen accessSenior author

  Automated detection of small and rare wildlife in aerial imagery is crucial for effective conservation, yet remains a significant technical challenge. Prairie dogs exemplify this issue: their ecological importance as keystone species contrasts sharply with their elusive presence--marked by small size, sparse distribution, and subtle visual features--which undermines existing detection approaches. To address these challenges, we propose RareSpot, a robust detection framework integrating multi-scale consistency learning and context-aware augmentation. Our multi-scale consistency approach leverages structured alignment across feature pyramids, enhancing fine-grained object representation and mitigating scale-related feature loss. Complementarily, context-aware augmentation strategically synthesizes challenging training instances by embedding difficult-to-detect samples into realistic environmental contexts, significantly boosting model precision and recall. Evaluated on an expert-annotated prairie dog drone imagery benchmark, our method achieves state-of-the-art performance, improving detection accuracy by over 35% compared to baseline methods. Importantly, it generalizes effectively across additional wildlife datasets, demonstrating broad applicability. The RareSpot benchmark and approach not only support critical ecological monitoring but also establish a new foundation for detecting small, rare species in complex aerial scenes.

  PublisherOA PDFDOI

• METAREG: Robust Camera Parameter Estimation by Leveraging Noisy Camera Extrinsics

  2025-08-18

  articleSenior author

  Novel view synthesis methods, such as Neural Radiance Fields (NeRFs) and 3D Gaussian Splatting (3DGS), rely on Structure-from-Motion (SfM) pipelines like COLMAP for camera parameter estimation. However, these pipelines are prone to errors due to factors like doppelgangers and perspective distortion. While edge devices (e.g., mobile phones) capture images with embedded GPS and IMU data, this meta-data is often noisy. We introduce MetaReg, a robust pipeline that improves camera parameter estimation by leveraging noisy GPS metadata. MetaReg enhances COLMAP with pre-and post-processing steps: the preprocessing stage estimates image overlap using metadata to reduce unnecessary image matching, while the post-processing stage aligns estimated camera coordinates with the world coordinate system using noisy GPS priors. Experiments on challenging datasets demonstrate that MetaReg significantly improves camera parameter estimation, enhancing robustness and accuracy.

  PublisherDOI

• Q-RBSA: high-resolution 3D EBSD map generation using an efficient quaternion transformer network

  npj Computational Materials · 2024-01-30 · 16 citations

  articleOpen accessSenior author

  Abstract Gathering 3D material microstructural information is time-consuming, expensive, and energy-intensive. Acquisition of 3D data has been accelerated by developments in serial sectioning instrument capabilities; however, for crystallographic information, the electron backscatter diffraction (EBSD) imaging modality remains rate limiting. We propose a physics-based efficient deep learning framework to reduce the time and cost of collecting 3D EBSD maps. Our framework uses a quaternion residual block self-attention network (QRBSA) to generate high-resolution 3D EBSD maps from sparsely sectioned EBSD maps. In QRBSA, quaternion-valued convolution effectively learns local relations in orientation space, while self-attention in the quaternion domain captures long-range correlations. We apply our framework to 3D data collected from commercially relevant titanium alloys, showing both qualitatively and quantitatively that our method can predict missing samples (EBSD information between sparsely sectioned mapping points) as compared to high-resolution ground truth 3D EBSD maps.

  PublisherOA PDFDOI

• Identification of Resistance Gene for Mungbean Yellow Mosaic Virus (MYMV) in Resistant Mungbean Genotypes through RGA Markers

  Journal of Advances in Biology & Biotechnology · 2024-05-30 · 1 citations

  articleOpen access

  Mungbean is one among the important and major pulse crop for supplementation of protein in subtropical zones of the world. The yield of mungbean is affected by both abiotic and biotic factors. Among the biotic factors, Mungbean yellow mosaic virus (MYMV) is the major bottleneck for pulse producers. MYMV cause a Yellow mosaic disease and its transmitted by Aleyrodidae insects. The virus belong to the family Geminiviridae and the genus Begomovirus. Fourteen mungbean genotypes procured from Asian Vegetable Research and Development Centre (AVRDC), Taiwan were screened for Mung bean Yellow Mosaic Virus resistance in field and laboratory conditions during Kharif 2022 and summer 2023. Genotypes MYB-6, 7, 8, 9 and 12 were found to be resistant under field conditions by less PDI and AUDPC values. The molecular confirmation of MYMV resistance was tested using Resistant Gene Analog (RGA) in all 14 genotypes. The RGA was amplified at 90bp in resistant genotypes viz MYB-6, 7, 8 and 9. Since MYMV resistant lines MYB-6, 7, 8 and 9 showed resistance in both field and molecular studies, hence these lines can be used as a YMV donor in breeding process and also can be released as a variety after conducting a multi-location trails in MYMV hotspot conditions.

  PublisherOA PDFDOI

• CIMGEN: Controlled Satellite Image Manipulation by Finetuning Pretrained Generative Models on Limited Data

  Lecture notes in computer science · 2024-12-03 · 4 citations

  book-chapterSenior author
  PublisherDOI

• Expression Study of Bmlipase in Silkworm Bombyx mori midgut against BmNPV

  Annual Research & Review in Biology · 2024-11-18 · 1 citations

  articleOpen access

  Silkworms are susceptible to various pathogens, including bacteria, fungi, viruses, and protozoa. The present study focused on the viral disease known as nucleopolyhedrosis virus, which causes severe infections in silkworms. Silkworms possess several antiviral proteins that play a crucial role in preventing the spread of infection. One protein, which is found in the digestive juice produced in the silkworm's midgut, was focused, and it is lipase. This protein is responsible for controlling the virus infection. To investigate the role of lipase expression in silkworm for virus control, its concentration after post-infection was measured. This was quantified and analysed using SDS-PAGE. The band of interest, with a molecular weight of 29 kDa, was further analysed through MALDI-TOF/MS, and it showed 70% homology to the lipase enzyme.

  PublisherOA PDFDOI

• Extracting Personality Characteristics from Handwriting using Machine Learning.

  International Journal For Multidisciplinary Research · 2024-09-09

  articleOpen accessSenior author

  Handwriting analysis has long been employed to glean insights into an individual's personality traits, behaviors, and psychological characteristics. From forensic handwriting analysis to psychological profiling, the intricate nuances of handwriting have been studied and interpreted by experts across various disciplines. With the advent of digital technologies and advancements in image processing techniques, the field of personality identification through handwriting analysis has witnessed a resurgence, offering new avenues for research and applications. This project endeavors to contribute to the ongoing discourse on personality identification through handwriting analysis by proposing a novel approach that leverages the power of image processing, specifically Convolutional Neural Networks (CNNs), to extract meaningful features from handwriting images and classify them based on personality traits.

  PublisherOA PDFDOI

• CIMGEN: Controlled Image Manipulation by Finetuning Pretrained Generative Models on Limited Data

  arXiv (Cornell University) · 2024-01-23

  preprintOpen accessSenior author

  Content creation and image editing can benefit from flexible user controls. A common intermediate representation for conditional image generation is a semantic map, that has information of objects present in the image. When compared to raw RGB pixels, the modification of semantic map is much easier. One can take a semantic map and easily modify the map to selectively insert, remove, or replace objects in the map. The method proposed in this paper takes in the modified semantic map and alter the original image in accordance to the modified map. The method leverages traditional pre-trained image-to-image translation GANs, such as CycleGAN or Pix2Pix GAN, that are fine-tuned on a limited dataset of reference images associated with the semantic maps. We discuss the qualitative and quantitative performance of our technique to illustrate its capacity and possible applications in the fields of image forgery and image editing. We also demonstrate the effectiveness of the proposed image forgery technique in thwarting the numerous deep learning-based image forensic techniques, highlighting the urgent need to develop robust and generalizable image forensic tools in the fight against the spread of fake media.

  PublisherOA PDFDOI

• ReeFRAME: Reeb Graph based Trajectory Analysis Framework to Capture Top-Down and Bottom-Up Patterns of Life

  2024-10-29 · 1 citations

  preprintOpen accessSenior author

  In this paper, we present ReeFRAME, a scalable Reeb graph-based framework designed to analyze vast volumes of GPS-enabled human trajectory data generated at 1Hz frequency. ReeFRAME models Patterns-of-life (PoL) at both the population and individual levels, utilizing Multi-Agent Reeb Graphs (MARGs) for population-level patterns and Temporal Reeb Graphs (TERGs) for individual trajectories. The framework's linear algorithmic complexity relative to the number of time points ensures scalability for anomaly detection. We validate ReeFRAME on six large-scale anomaly detection datasets, simulating real-time patterns with up to 500,000 agents over two months.

  PublisherOA PDFDOI

Recent grants

• The connectome and neurobiology of a novel model chordate, Ciona intestinalis

  NIH · $2.2M · 2017–2023

• Information Technology Research (ITR): Next-Generation Bio-Molecular Imaging and Information Discovery

  NSF · $7.7M · 2003–2010

• The connectome and neurobiology of a novel model chordate, Ciona intestinalis

  NIH · $538k · 2017–2022

• Morphomic analysis of a simple chordate

  NIH · $2.5M · 2009–2018

• CDI-Type-II: Computational Challenges in the Discovery and Understanding of Complex Boiological Structures through Multimodal Imaging

  NSF · $1.9M · 2009–2013

Frequent coauthors

• Shivkumar Chandrasekaran

  72 shared

• Lakshmanan Nataraj

  63 shared

• Tajuddin Manhar Mohammed

  Mayachitra (United States)

  39 shared

• Charles Kenney

  26 shared

• Jawadul H. Bappy

  20 shared

• Michael Goebel

  20 shared

• Baris Sumengen

  Google (United States)

  20 shared

• Upamanyu Madhow

  University of California, Santa Barbara

  19 shared

Labs

• Vision Research LabPI

  Principal investigator, researchers, students and staff.