Searching for Ambiguous Objects in Videos using Relational Referring Expressions

Date

2019-09-12

Author

Anayurt, Hazan
Özyeğin, Sezai Artun
Çetin, Ülfet
Aktaş, Utku
Kalkan, Sinan

Metadata

Show full item record

Item Usage Stats

142
views

0
downloads

Humans frequently use referring (identifying) expressions to refer to objects. Especially in ambiguous settings, humans prefer expressions (called relational referring expressions) that describe an object with respect to a distinguishing, unique object. Unlike studies on video object search using referring expressions, in this paper, our focus is on (i) relational referring expressions in highly ambiguous settings, and (ii) methods that can both generate and comprehend a referring expression. For this goal, we first introduce a new dataset for video object search with referring expressions that includes numerous copies of the objects, making it difficult to use non-relational expressions. Moreover, we train two baseline deep networks on this dataset, which show promising results. Finally, we propose a deep attention network that significantly outperforms the baselines on our dataset

URI

https://hdl.handle.net/11511/76009

Conference Name

30th British Machine Vision Conference (BMVC), (9-12 Eylül 2019)

Collections

Department of Computer Engineering, Conference / Seminar

Suggestions

OpenMETU
Core

Learning to Generate Unambiguous Spatial Referring Expressions for Real-World Environments Dagan, Fethiye Irmak; Kalkan, Sinan; Leite, Iolanda (2019-01-01) Referring to objects in a natural and unambiguous manner is crucial for effective human-robot interaction. Previous research on learning-based referring expressions has focused primarily on comprehension tasks, while generating referring expressions is still mostly limited to rule-based methods. In this work, we propose a two-stage approach that relies on deep learning for estimating spatial relations to describe an object naturally and unambiguously with a referring expression. We compare our method to the...
Investigating the Neural Models for Irony Detection on Turkish Informal Texts Cemek, Yesim; Cidecio, Cenk; Ozturk, Asli Umay; Cekinel, Recep Firat; Karagöz, Pınar (2020-01-01) Irony is defined as the expression of one's meaning by using language that normally signifies the opposite, typically for humorous or emphatic effect. In the textual context, it can be considered as a specific type of opinion mining problem. However, due to the nature of the problem, it is generally more challenging to detect. Irony detection is useful to understand the semantics of a text. It also helps improving the accuracy of many other text mining tasks. In this paper, we study the irony detection prob...
Employing Named Entities for Semantic Retrieval of News Videos in Turkish Kucuk, Dilek; Yazıcı, Adnan (2009-09-16) Named entities are known to be important means for semantic annotation of news texts. Considerable work has been carried out for semantic indexing of both textual news and news videos especially in English through the employment of named entities extracted from textual news or transcriptions of the news videos. In this paper, we present our semantic retrieval architecture for news videos in Turkish based on prior semantic annotation of the videos with the corresponding named entities in the news transcripti...
Identification of factors affecting the e-service adoption: an empirical investigation / Çetin Kaya, Yasemin; Özkan Yıldırım, Sevgi; Department of Information Systems (2014) E-service usage has become widespread and constitutes a significant place in the daily life of people. Despite the diversity of benefits offered by e-service, adoption problems cause reduced exploitation of e-services. In this regard, in order to extend the use of e-services it has become essential to determine the needs and expectations of users. The aim of this study is to develop and validate an e-service adoption model that comprises determinants of behavioral, normative and control belief structures. A...
INVESTIGATING THE ASYMMETRIC NATURE OF THE CONTIGUITY EFFECT VIA PROBED RECALL TASK ARPACI, Hazal; KILIÇ ÖZHAN, Aslı; Department of Psychology (2022-7-26) In free recall, there is a tendency to generate a word that either follows or precedes the just recalled word in the study list, which is known as the contiguity effect. This effect has been explained by two main accounts: causal models and non-causal models. Causal models claim that the contiguity effect occurs due to the utilization of the just recalled item as a cue to recall the next item, whereas according to non-causal models, items are not used as cues, but instead the similarity between the mental s...

Citation Formats

H. Anayurt, S. A. Özyeğin, Ü. Çetin, U. Aktaş, and S. Kalkan, “Searching for Ambiguous Objects in Videos using Relational Referring Expressions,” presented at the 30th British Machine Vision Conference (BMVC), (9-12 Eylül 2019), 2019, Accessed: 00, 2021. [Online]. Available: https://hdl.handle.net/11511/76009.