1 code implementation • 14 Mar 2024 • Vibashan VS, Shubhankar Borse, Hyojin Park, Debasmit Das, Vishal Patel, Munawar Hayat, Fatih Porikli
In this paper, we introduce an open-vocabulary panoptic segmentation model that effectively unifies the strengths of the Segment Anything Model (SAM) with the vision-language CLIP model in an end-to-end framework.
Ranked #1 on Open Vocabulary Panoptic Segmentation on ADE20K
Open Vocabulary Panoptic Segmentation Panoptic Segmentation +1
no code implementations • ICCV 2023 • Jungsoo Lee, Debasmit Das, Jaegul Choo, Sungha Choi
To be more specific, entropy minimization attempts to raise the confidence values of an individual sample's prediction, but individual confidence values may rise or fall due to the influence of signals from numerous other predictions (i. e., wisdom of crowds).
no code implementations • CVPR 2023 • Seokeon Choi, Debasmit Das, Sungha Choi, Seunghan Yang, Hyunsin Park, Sungrack Yun
Single domain generalization aims to train a generalizable model with only one source domain to perform well on arbitrary unseen target domains.
no code implementations • CVPR 2023 • Shubhankar Borse, Debasmit Das, Hyojin Park, Hong Cai, Risheek Garrepalli, Fatih Porikli
Next, we use a conditional regenerator, which takes the redacted image and the dense predictions as inputs, and reconstructs the original image by filling in the missing structural information.
no code implementations • 24 Feb 2023 • Debasmit Das, Shubhankar Borse, Hyojin Park, Kambiz Azarian, Hong Cai, Risheek Garrepalli, Fatih Porikli
Test-time adaptive (TTA) semantic segmentation adapts a source pre-trained image semantic segmentation model to unlabeled batches of target domain test images, different from real-world, where samples arrive one-by-one in an online fashion.
no code implementations • 12 Dec 2022 • Kambiz Azarian, Debasmit Das, Hyojin Park, Fatih Porikli
In this approach, we do not assume test-time access to the labeled source dataset.
1 code implementation • 4 Jul 2022 • Davide Belli, Debasmit Das, Bence Major, Fatih Porikli
Face authentication systems require a robust anti-spoofing module as they can be deceived by fabricating spoof images of authorized users.
Ranked #1 on Face Anti-Spoofing on SiW (Protocol 3)
no code implementations • 28 Jun 2022 • Seunghan Yang, Debasmit Das, Janghoon Cho, Hyoungwoo Park, Sungrack Yun
Deep learning models for verification systems often fail to generalize to new users and new environments, even though they learn highly discriminative features.
no code implementations • 24 Jun 2022 • Jihoon Moon, Debasmit Das, C. S. George Lee
To project the data from the source and the target domains to a common subspace and manipulate the projected data in real-time, our proposed framework institutes a novel method, called an Incremental Computation of Mean-Subspace (ICMS) technique, which computes an approximation of mean-target subspace on a Grassmann manifold and is proven to be a close approximate to the Karcher mean.
no code implementations • CVPR 2022 • Shubhankar Borse, Hyojin Park, Hong Cai, Debasmit Das, Risheek Garrepalli, Fatih Porikli
A Panoptic Relational Attention (PRA) module is then applied to the encodings and the global feature map from the backbone.
1 code implementation • WACV 2022 • Davide Belli, Debasmit Das, Bence Major, Fatih Porikli
In real-world scenarios, however, face authentication systems often have an initial enrollment step where a few live images of the user are recorded and stored for identification purposes.
Ranked #1 on Face Anti-Spoofing on CelebA-Spoof-Enroll5
no code implementations • 24 Nov 2021 • Seunghan Yang, Debasmit Das, Simyung Chang, Sungrack Yun, Fatih Porikli
However, it is observed that image transformations already present in the dataset might be less effective in learning such self-supervised representations.
no code implementations • ICLR 2022 • Debasmit Das, Sungrack Yun, Fatih Porikli
The first step of our framework trains a feature extracting backbone with the contrastive loss on the base category data.
no code implementations • 2 May 2021 • Debasmit Das, Yash Bhalgat, Fatih Porikli
The initialization is cast as an optimization problem where we minimize a combination of encoding and decoding losses of the input activations, which is further constrained by a user-defined latent code.
no code implementations • 22 Oct 2020 • Debasmit Das, J. H. Moon, C. S. George Lee
In this paper, we extend the traditional few-shot learning (FSL) problem to the situation when the source-domain data is not accessible but only high-level information in the form of class prototypes is available.
no code implementations • 20 Feb 2020 • J. H. Moon, Debasmit Das, C. S. George Lee
The traditional methods on the OUDA problem mainly focus on transforming each arriving target data to the source domain, and they do not sufficiently consider the temporal coherency and accumulative statistics among the arriving target data.
no code implementations • 10 Dec 2019 • Debasmit Das, C. S. George Lee
This paper proposes a multi-layer neural network structure for few-shot image recognition of novel categories.
no code implementations • 27 Mar 2019 • Debasmit Das, C. S. George Lee
Secondly, we propose test-time domain adaptation to adapt the semantic embedding of the unseen classes to the test data.
no code implementations • 22 May 2018 • Debasmit Das, C. S. George Lee
Domain adaptation (DA) addresses the real-world image classification problem of discrepancy between training (source) and testing (target) data distributions.
no code implementations • 1 May 2018 • Debasmit Das, C. S. George Lee
The procedure of tackling this discrepancy between the training (source) and testing (target) domains is known as domain adaptation.