no code implementations • 11 Nov 2023 • Souradeep Chakraborty, Shujon Naha, Muhammet Bastan, Amit Kumar K C, Dimitris Samaras
Our unsupervised model is a great pre-training initialization for our semi-supervised model SS-CoSOD, especially when very limited labeled data is available for training.
no code implementations • 1 Nov 2023 • Oriol Barbany, Xiaofan Lin, Muhammet Bastan, Arnab Dhua
Deep Metric Learning (DML) methods aim at learning an embedding space in which distances are closely related to the inherent semantic similarity of the inputs.
no code implementations • 25 Mar 2021 • Zhibo Yang, Muhammet Bastan, Xinliang Zhu, Doug Gray, Dimitris Samaras
In this paper, we present a framework that leverages this implicit hierarchy by imposing a hierarchical structure on the proxies and can be used with any existing proxy-based loss.
no code implementations • 17 May 2020 • Muhammet Bastan, Arnau Ramisa, Mehmet Tek
Transformer models have recently achieved impressive performance on NLP tasks, owing to new algorithms for self-supervised pre-training on very large text corpora.
1 code implementation • 18 Nov 2019 • Muhammet Bastan, Hao-Yu Wu, Tian Cao, Bhargava Kota, Mehmet Tek
We present an open-set logo detection (OSLD) system, which can detect (localize and recognize) any number of unseen logo classes without re-training; it only requires a small set of canonical logo images for each logo class.
no code implementations • 14 Nov 2019 • Dipu Manandhar, Muhammet Bastan, Kim-Hui Yap
In view of this, we propose a new deep semantic granularity metric learning (SGML) that develops a novel idea of leveraging attribute semantic space to capture different granularity of similarity, and then integrate this information into deep metric learning.
no code implementations • 28 Apr 2018 • Muhammet Bastan, Kim-Hui Yap, Lap-Pui Chau
First, we detect the cars in each IR image using a convolutional neural network, which is pre-trained on regular RGB images and fine-tuned on IR images for higher accuracy.
no code implementations • 12 Sep 2016 • Muhammet Bastan, S. Saqib Bukhari, Thomas M. Breuel
This way, the strong edges are used to recover weak or missing edges by considering the local edge structures, instead of blindly linking them if gradient magnitudes are above some threshold.
no code implementations • 11 Aug 2016 • Muhammet Bastan, Ozgur Yilmaz
We concluded that (1) multi-view queries with deep ConvNets representations perform significantly better than single view queries, (2) ConvNets perform much better than BoWs and have room for further improvement, (3) pre-training of ConvNets on a different image dataset with background clutter is needed to obtain good performance on cluttered product image queries obtained with a mobile phone.
no code implementations • 31 Jul 2015 • Fatih Calisir, Muhammet Bastan, Ozgur Ulusoy, Ugur Gudukbay
High user interaction capability of mobile devices can help improve the accuracy of mobile visual search systems.