About

I am a Ph.D. candidate at Indiana University Bloomington. I work in Computer Vision Lab with Prof. David Crandall. I am majoring in Computer Science with a minor in Machine Learning. Before joining the Ph.D. program, I earned a master’s degree in Data Science at Indiana University with a thesis on ‘Response-Based Knowledge Distillation’. I completed a Bachelor’s degree in Electrical Engineering at the National Institute of Technology, Patna (NITP).

(Résumé).

Research Interests

My research involves developing and applying computer vision algorithms that understand, analyze, and organize large-scale datasets of images and video using deep learning-based techniques. My recent focus has been on creating computer vision algorithms that can generate and synthesize 3D models of visual scenes from 2D images, using techniques such as multi-view stereo, stable diffusion, and other deep learning models

• 3D reconstruction with learning-based Multi-View Stereo
• Geometrical reasoning of 3D scenes
• Deep Learning and Case-Based Reasoning (DL-CBR) integration
• Deep Learning - Architectural understanding and exploration
• Generative models for geospatial images

Publications

Journal Articles

• Vibhas Vats and D. Crandall, “Geometric Constraints in Deep Learning Frameworks: A Survey” Just Accepted at ACM-surveys, 2025. ACM-paper (PrePrint)
• Cuhua Wang, Md. Reza, Vibhas Vats, Y. Ju, N. Thankurdesai, Y. Wang, D. Crandall, J. Seo, Soon-Heung Jung, “Deep Learning-based 3D Reconstruction from Multiple Images: A Survey”, Neurocomputing 2024 (Paper)

Conference Proceedings

• Vibhas Kumar Vats and Z. Wilkerson and H. Sato and David Leake and David Crandall,”Learning Case Features with Proxy-Guided Deep Neural Networks”, Accepted at ICCBR 2025
• Vibhas Vats, Sripad Joshi, David Crandall, Md. Reza, Soon-Heung Jung, “GC-MVSNet: Multi-View, Multi-Scale, Geometrically-Consistent Multi-View Stereo”, WACV 2024 (Project Page)
• Vibhas Vats, Zachary Wilkerson, David Leake, David Crandall, “Extracting Indexing Features for CBR from Deep Neural Networks: A Transfer Learning Approach”, ICCBR 2024 Paper
• Zachary Wilkerson, Vibhas Vats, Karan Acharya, David Leake, David Crandall, “Examining the Impact of Network Architecture on Extracted Feature Quality for CBR” ICCBR-2023 (Paper)
• Vibhas Vats, David Crandall (2021). “Controlling the Quality of Distillation in Response-Based Network Compression.” AAAI-22. 1-8. (Paper)
• VK Vats, S Rai, S De, M De (2018). “Mitigating Effect of Communication Link Failure in Smart Meter-Based Load Forecasting.” Springer. 289-300. (Paper)
• VK Vats, S Rai, D Bharti, Mala De. (2018). “Very Short-term, Short-Term and Mid-Term Load Forecasting for Residential Academic Institute: A Case Study.” IEEE. 1-6. (Paper)

PrePrint

• Vibhas Vats and David Crandall, “Geometric Constraints in Deep Learning Frameworks: A Survey” 2024. (Paper)

Under-review

• Vibhas Vats, M. A. Reza, D. J. Crandall, and S.-h. Jung, “Blending 3D Geometry and Machine Learning for Multi-View Stereopsis” (Under review - TPAMI), Nov-2024.
• Tony Ha, V. K. Vats, M. A. R. Soon-heung Jung, and D. Crandall, “Hvpunet: Hybrid-voxel point-cloud unsampling network” (under review - ICCV), 2025.

Patents

• Soon-Heung Jung, David Crandall, Vibhas Kumar Vats, Shaurya Shubham, Md Alimoor Reza, Chuhua Wang, “Method and Apparatus for Estimating Depth Information of Images”, USA Google Patents link
• Soon-Heung Jung, Vibhas Kumar Vats, David J Crandall, Md Alimoor Reza, Sripad Joshi, “Learning method and device for estimating depth information of image”, USA, Google Patents Link

Master’s Thesis

• Response-Based Knowledge Distillation. (pdf)

Please see my Google Scholar Profile

Teaching

I co-teach Computer Vision (CSCI-B 657) with my advisor every Spring since 2022. I discuss a number of seminal papers exploring deep learning (DL) architectures. I designed the first iteration of the DL discussion section in Spring 2022, it is designed to explore the seminal papers exploring DL architectures to develop an intuitive as well as mathematical understanding of major concepts in DL. In 16 weeks, the course roughly covers major architectures in three broad sections, Convolution-based networks (CNNs), Multi-layer perceptron-based networks (MLPs), and Transformer-based networks (ViTs). In CNNs, we cover models like LeNet, AlexNet, GoogLeNet, ResNets, Wide ResNet, Stochastic ResNet, ResNeXt, DenseNet, ConvMixer, and Xception Net. In MLPs, we cover models like MLP Mixer, ResMLP, cycleMLP, and S^2 MLP model, and in ViTs, we cover models like Vision Transformer, DeiT, Swin Transformer, Local ViT, Convolutional-ViT, etc. We also explore shift operation-based networks. The class meets weekly with more than 150 students participating in the discussion. The list of papers covered in the discussion is below

Preprints

Preprints are available on arXiv

Service

Reviewer @

• International Conference on Intelligence Science (ICIS)
• IET image processing
• International Conference on Intelligent Robots and Systems (IROS)
• IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
• IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)

Fun

I run for fun :). Find me on the Nike Run Club (@Vibhas_Vats). Hiking, driving, reading, and writing are calming things to do.