no code implementations • 8 Nov 2023 • Muhammad Akhtar Munir, Muhammad Haris Khan, M. Saquib Sarfraz, Mohsen Ali
Deep learning based object detectors struggle generalizing to a new target domain bearing significant variations in object and background.
1 code implementation • NeurIPS 2023 • Muhammad Akhtar Munir, Salman Khan, Muhammad Haris Khan, Mohsen Ali, Fahad Shahbaz Khan
Third, we develop a logit mixing approach that acts as a regularizer with detection-specific losses and is also complementary to the uncertainty-guided logit modulation technique to further improve the calibration performance.
no code implementations • 27 Sep 2023 • Javed Iqbal, Aliza Masood, Waqas Sultani, Mohsen Ali
In this work, we propose a topology-aware unsupervised domain adaptation approach for road segmentation in remote sensing imagery.
no code implementations • 17 Sep 2023 • Arif Mahmood, Abdul Basit, M. Akhtar Munir, Mohsen Ali
By combining these components, our approach achieves exceptional results on a newly proposed dataset.
no code implementations • 15 Sep 2022 • Muhammad Akhtar Munir, Muhammad Haris Khan, M. Saquib Sarfraz, Mohsen Ali
To this end, we first propose a new, plug-and-play, train-time calibration loss for object detection (coined as TCD).
no code implementations • 20 Jun 2022 • Javed Iqbal, Hamza Rawal, Rehan Hafiz, Yu-Tseh Chi, Mohsen Ali
Due to the domain shift, this decision boundary is unaligned in the target domain, resulting in noisy pseudo labels adversely affecting self-supervised domain adaptation.
no code implementations • 9 Apr 2022 • M. Fasi ur Rehman, Izza Ali, Waqas Sultani, Mohsen Ali
A small set of seed samples (32 in our case) are automatically discovered by analyzing the temporal changes, which are manually labeled to train a segmentation and representation learning module.
no code implementations • 7 Jan 2022 • Javed Iqbal, Rehan Hafiz, Mohsen Ali
We propose a self-entropy and multi-scale information augmented self-supervised domain adaptation method (FogAdapt) to minimize the domain shift in foggy scenes segmentation.
Ranked #1 on Domain Adaptation on SYNTHIA-to-FoggyCityscapes
no code implementations • NeurIPS 2021 • Muhammad Akhtar Munir, Muhammad Haris Khan, M. Sarfraz, Mohsen Ali
In this paper, we propose to leverage model’s predictive uncertainty to strike the right balance between adversarial feature alignment and class-level alignment.
no code implementations • CVPR 2022 • Waqas Sultani, Wajahat Nawaz, Syed Javed, Muhammad Sohail Danish, Asma Saadia, Mohsen Ali
We design a mechanism to transfer these annotations from the high-cost microscope at high magnification to the low-cost microscope, at multiple magnifications.
no code implementations • 2 Nov 2021 • Muhammad Zaida, Shafaqat Ali, Mohsen Ali, Sarfaraz Hussein, Asma Saadia, Waqas Sultani
Deep neural networks have shown promising results in disease detection and classification using medical image data.
1 code implementation • 26 Oct 2021 • Muaaz Zakria, Hamza Rawal, Waqas Sultani, Mohsen Ali
We then exploit counting consistency constraints, within-image count consistency, and across-image count consistency, to decrease the domain shift.
no code implementations • 1 Oct 2021 • Muhammad Akhtar Munir, Muhammad Haris Khan, M. Saquib Sarfraz, Mohsen Ali
In this paper, we propose to leverage model predictive uncertainty to strike the right balance between adversarial feature alignment and class-level alignment.
no code implementations • 17 Feb 2021 • Qazi Ammar Arshad, Mohsen Ali, Saeed-Ul Hassan, Chen Chen, Ayisha Imran, Ghulam Rasul, Waqas Sultani
Malaria microscopy, microscopic examination of stained blood slides to detect parasite Plasmodium, is considered to be a gold-standard for detecting life-threatening disease malaria.
1 code implementation • ECCV 2020 • M. Naseer Subhani, Mohsen Ali
Specifically, we show that semantic segmentation model produces output with high entropy when presented with scaled-up patches of target domain, in comparison to when presented original size images.
1 code implementation • 5 Jul 2020 • Javed Iqbal, Mohsen Ali
We thoroughly study the limitations of existing domain adaptation methods and propose a weakly-supervised adaptation strategy where we assume image-level labels are available for the target domain.
no code implementations • 19 May 2020 • Abdul Basit, Muhammad Akhtar Munir, Mohsen Ali, Arif Mahmood
Visual identification of gunmen in a crowd is a challenging problem, that requires resolving the association of a person with an object (firearm).
1 code implementation • ISPRS Journal of Photogrammetry and Remote Sensing 2020 • Muhammad Usman Ali, Waqas Sultani, Mohsen Ali
Natural and man-made disasters cause huge damage to built infrastructures and results in loss of human lives.
1 code implementation • 30 Sep 2019 • Javed Iqbal, Mohsen Ali
Thus helping latent space learn the representation even when there are very few pixels belonging to the domain category (small object for example) compared to rest of the image.
3 code implementations • WACV 2020 • Muhammad Faisal, Ijaz Akhter, Mohsen Ali, Richard Hartley
To handle the nonrigid background like a sea, we also propose a robust fusion mechanism between motion and appearance-based features.
no code implementations • IEEE Access 2019 • Aman Irshad, Rehan Hafiz, Mohsen Ali, Muhammad Faisal, Yongju Cho, Jeongil Seo
Our results on Brown and HPatches datasets demonstrate Twin-Net's consistently better performance as well as better discriminatory and generalization capability as compared to the state-of-art.
Ranked #1 on Patch Matching on HPatches
1 code implementation • 22 Apr 2019 • Javed Iqbal, Muhammad Akhtar Munir, Arif Mahmood, Afsheen Rafaqat Ali, Mohsen Ali
The OAOD algorithm is evaluated on the ITUF dataset and compared with current state-of-the-art object detectors, including fully supervised oriented object detectors.
no code implementations • 1 Apr 2019 • Anza Shakeel, Waqas Sultani, Mohsen Ali
In this paper, we attempt to address the challenging problem of counting built-structures in the satellite imagery.
1 code implementation • 25 Jul 2017 • Afsheen Rafaqat Ali, Mohsen Ali
In this paper we present an automatic image-transformation method that transforms the source image such that it can induce an emotional affect on the viewer, as desired by the user.
no code implementations • 12 May 2017 • Anza Shakeel, Mohsen Ali
Deep convolutional neural networks (CNNs) have outperformed existing object recognition and detection algorithms.
no code implementations • 8 May 2017 • Afsheen Rafaqat Ali, Usman Shahid, Mohsen Ali, Jeffrey Ho
This paper aims to bridge the affective gap between image content and the emotional response of the viewer it elicits by using High-Level Concepts (HLCs).
no code implementations • CVPR 2013 • Yu-Tseh Chi, Mohsen Ali, Ajit Rajwade, Jeffrey Ho
This paper proposes a dictionary learning framework that combines the proposed block/group (BGSC) or reconstructed block/group (R-BGSC) sparse coding schemes with the novel Intra-block Coherence Suppression Dictionary Learning (ICS-DL) algorithm.