Subhransu Maji

Affiliations

University of Massachusetts Amherst

papers

1,035

total citations

papers (21)

ParSeNet: A Parametric Surface Fitting Network for 3D Point Clouds

ECCV 2020arXiv

189

citations

When Does Self-supervision Improve Few-shot Learning?

ECCV 2020arXiv

188

citations

Exponential Moving Average Normalization for Self-Supervised and Semi-Supervised Learning

CVPR 2021arXiv

148

citations

PhraseCut: Language-Based Image Segmentation in the Wild

CVPR 2020arXiv

138

citations

Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions

CVPR 2024arXiv

citations

A Realistic Evaluation of Semi-Supervised Learning for Fine-Grained Classification

CVPR 2021arXiv

citations

Label-Efficient Learning on Point Clouds using Approximate Convex Decompositions

ECCV 2020arXiv

citations

MvDeCor: Multi-View Dense Correspondence Learning for Fine-Grained 3D Segmentation

ECCV 2022arXiv

citations

WildSAT: Learning Satellite Image Representations from Wildlife Observations

ICCV 2025arXiv

citations

On Equivariant and Invariant Learning of Object Landmark Representations

ICCV 2021arXiv

citations

Human-in-the-Loop Visual Re-ID for Population Size Estimation

ECCV 2024arXiv

citations

Consensus-Driven Active Model Selection

ICCV 2025arXiv

citations

Generate, Transduct, Adapt: Iterative Transduction with VLMs

ICCV 2025arXiv

citations

Merlin L48 Spectrogram Dataset

NEURIPS 2025arXiv

citations

Task2Box: Box Embeddings for Modeling Asymmetric Task Relationships

CVPR 2024arXiv

citations

Subhransu Maji

Affiliations

papers (21)

ParSeNet: A Parametric Surface Fitting Network for 3D Point Clouds

When Does Self-supervision Improve Few-shot Learning?

Exponential Moving Average Normalization for Self-Supervised and Semi-Supervised Learning

PhraseCut: Language-Based Image Segmentation in the Wild

Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions

A Realistic Evaluation of Semi-Supervised Learning for Fine-Grained Classification

Label-Efficient Learning on Point Clouds using Approximate Convex Decompositions

LU-NeRF: Scene and Pose Estimation by Synchronizing Local Unposed NeRFs

Cross-Modal 3D Shape Generation and Manipulation

Learning Generative Models of Shape Handles

GanOrCon: Are Generative Models Useful for Few-Shot Segmentation?

Improving Few-Shot Part Segmentation Using Coarse Supervision

Describing Textures using Natural Language

MvDeCor: Multi-View Dense Correspondence Learning for Fine-Grained 3D Segmentation

WildSAT: Learning Satellite Image Representations from Wildlife Observations

On Equivariant and Invariant Learning of Object Landmark Representations

Human-in-the-Loop Visual Re-ID for Population Size Estimation

Consensus-Driven Active Model Selection

Generate, Transduct, Adapt: Iterative Transduction with VLMs

Merlin L48 Spectrogram Dataset

Task2Box: Box Embeddings for Modeling Asymmetric Task Relationships

papers (21)

ParSeNet: A Parametric Surface Fitting Network for 3D Point Clouds

When Does Self-supervision Improve Few-shot Learning?

Exponential Moving Average Normalization for Self-Supervised and Semi-Supervised Learning

PhraseCut: Language-Based Image Segmentation in the Wild

Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions

A Realistic Evaluation of Semi-Supervised Learning for Fine-Grained Classification

Label-Efficient Learning on Point Clouds using Approximate Convex Decompositions

LU-NeRF: Scene and Pose Estimation by Synchronizing Local Unposed NeRFs

Cross-Modal 3D Shape Generation and Manipulation

Learning Generative Models of Shape Handles

GanOrCon: Are Generative Models Useful for Few-Shot Segmentation?

Improving Few-Shot Part Segmentation Using Coarse Supervision

Describing Textures using Natural Language

MvDeCor: Multi-View Dense Correspondence Learning for Fine-Grained 3D Segmentation

WildSAT: Learning Satellite Image Representations from Wildlife Observations

On Equivariant and Invariant Learning of Object Landmark Representations

Human-in-the-Loop Visual Re-ID for Population Size Estimation

Consensus-Driven Active Model Selection

Generate, Transduct, Adapt: Iterative Transduction with VLMs

Merlin L48 Spectrogram Dataset

Task2Box: Box Embeddings for Modeling Asymmetric Task Relationships