Search Results for author: Ashraful Islam

Found 8 papers, 5 papers with code

Self-supervised Learning with Local Contrastive Loss for Detection and Semantic Segmentation

no code implementations10 Jul 2022 Ashraful Islam, Ben Lundell, Harpreet Sawhney, Sudipta Sinha, Peter Morales, Richard J. Radke

We evaluate our SSL approach on two downstream tasks -- object detection and semantic segmentation, using COCO, PASCAL VOC, and CityScapes datasets.

object-detection Object Detection +2

ConfLab: A Rich Multimodal Multisensor Dataset of Free-Standing Social Interactions in the Wild

1 code implementation10 May 2022 Chirag Raman, Jose Vargas-Quiros, Stephanie Tan, Ekin Gedik, Ashraful Islam, Hayley Hung

Recording the dynamics of unscripted human interactions in the wild is challenging due to the delicate trade-offs between several factors: participant privacy, ecological validity, data fidelity, and logistical overheads.

Privacy Preserving

Generating Cyber Threat Intelligence to Discover Potential Security Threats Using Classification and Topic Modeling

no code implementations16 Aug 2021 Md Imran Hossen, Ashraful Islam, Farzana Anowar, Eshtiak Ahmed, Mohammad Masudur Rahman

In this paper, we identify and explore relevant CTI from hacker forums utilizing different supervised (classification) and unsupervised learning (topic modeling) techniques.

Computer Security

Dynamic Distillation Network for Cross-Domain Few-Shot Recognition with Unlabeled Data

1 code implementation NeurIPS 2021 Ashraful Islam, Chun-Fu Chen, Rameswar Panda, Leonid Karlinsky, Rogerio Feris, Richard J. Radke

As the base dataset and unlabeled dataset are from different domains, projecting the target images in the class-domain of the base dataset with a fixed pretrained model might be sub-optimal.

cross-domain few-shot learning

A Hybrid Attention Mechanism for Weakly-Supervised Temporal Action Localization

1 code implementation3 Jan 2021 Ashraful Islam, Chengjiang Long, Richard Radke

Moreover, our temporal semi-soft and hard attention modules, calculating two attention scores for each video snippet, help to focus on the less discriminative frames of an action to capture the full action boundary.

Hard Attention Multiple Instance Learning +2

DOA-GAN: Dual-Order Attentive Generative Adversarial Network for Image Copy-Move Forgery Detection and Localization

no code implementations CVPR 2020 Ashraful Islam, Chengjiang Long, Arslan Basharat, Anthony Hoogs

To the best of our knowledge, we are the first to propose such a network architecture with the 1st-order attention mechanism from the affinity matrix.

Weakly Supervised Temporal Action Localization Using Deep Metric Learning

1 code implementation21 Jan 2020 Ashraful Islam, Richard J. Radke

We propose a classification module to generate action labels for each segment in the video, and a deep metric learning module to learn the similarity between different action instances.

Metric Learning Temporal Localization +3

Cannot find the paper you are looking for? You can Submit a new open access paper.