Knowledge based query expansion in complex multimedia event detection

Boer, M.H.T. de; Schutte, Klamer; Kraaij, Wessel

doi:10.1007/s11042-015-2757-4

Cited by 22 publications

(8 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, "music" is a context word added to evaluate the term-term association of "information" and "technology" in the context of the query "music technology". Such context words can be extracted from a corpus using term co-occurrence [20,21,271,132] or derived from logical significance of a knowledge base [157,79,162,44].…”

Section: One-to-many Associationmentioning

confidence: 99%

“…Some other recent applications of QE are plagiarism detection [203], event search [89,15,43], text classification [269], patent retrieval [180,181,268], dynamic process in IoT [122,123], classification of e-commerce [128], biomedical IR [1], enterprise search [174], code search [205], parallel computing in IR [179] and twitter search [151,304]. Table 7 summarizes some of the prominent and recent applications of QE in literature based on the above discussion.…”

Section: Other Applicationsmentioning

confidence: 99%

See 1 more Smart Citation

Query expansion techniques for information retrieval: A survey

Azad

Deepak

2019

Information Processing & Management

204

104

View full text Add to dashboard Cite

With the ever increasing size of the web, relevant information extraction on the Internet with a query formed by a few keywords has become a big challenge. Query Expansion (QE) plays a crucial role in improving searches on the Internet. Here, the user's initial query is reformulated by adding additional meaningful terms with similar significance. QE -as part of information retrieval (IR) -has long attracted researchers' attention. It has become very influential in the field of personalized social document, question answering, cross-language IR, information filtering and multimedia IR. Research in QE has gained further prominence because of IR dedicated conferences such as TREC (Text Information Retrieval Conference) and CLEF (Conference and Labs of the Evaluation Forum). This paper surveys QE techniques in IR from 1960 to 2017 with respect to core techniques, data sources used, weighting and ranking methodologies, user participation and applications -bringing out similarities and differences.There is a huge amount of data available on the Internet, and it is growing exponentially. This unconstrained information-growth has not been accompanied by a corresponding technical advancement in the approaches for extracting relevant information [191]. Often, a web-search does not yield relevant results. There are multiple reasons for this. First, the keywords submitted by the user can be related to multiple topics; as a result, the search results are not focused on the topic of interest. Second, the query can be too short to capture appropriately what the user is looking for. This can happen just as a matter of habit (e.g., the average size of a web search is 2.4 words [257,255]). Third, the user is often not sure about what he is looking for until he sees the results. Even if the user knows what he is searching for, he does not know how to formulate an appropriate query (navigational queries are exceptions to this [51]). QE plays an important part in fetching relevant results in the above cases. Most web queries fall under the following three fundamental categories [51, 139] :-Informational Queries: Queries that cover a broad topic (e.g., India or journals) for which there may be thousands of relevant results. -Navigational Queries: Queries that are looking for specific website or URL (e.g., ISRO).-Transactional Queries: Queries that demonstrate the user's intent to execute a specific activity (e.g., downloading papers or buying books).2 Hiteshwar Kumar Azad, Akshay DeepakCurrently, user-queries are mostly processed using indexes and ontologies, which work on exact matches and are hidden from the users. This leads to the problem of term mismatch: user queries and search index are not based on the same set of terms. This is also known as the vocabulary problem [99]; it results from a combination of synonymy and polysemy. Synonymy refers to multiple words with common meaning, e.g., "buy" and "purchase". Polysemy refers to words with multiple meanings, e.g., "mouse" (a computer device or an animal). Synonymous and polysemou...

show abstract

Section: One-to-many Associationmentioning

confidence: 99%

Section: Other Applicationsmentioning

confidence: 99%

Query expansion techniques for information retrieval: A survey

Azad

Deepak

2019

Information Processing & Management

204

104

View full text Add to dashboard Cite

show abstract

“…To improve semantic matching, Alexiou et al (2016) correct class names to increase unseen action discrimination. Similar in spirit are approaches that employ query expansion (Dalton et al 2013;de Boer et al 2016) or textual action descriptions (Gan et al (2016c); Habibian et al 2017;Wang and Chen 2017) to make the action inputs more expressive. In contrast, we focus on improving the semantic matching itself to deal with semantic ambiguity, non-discriminative objects, and object naming.…”

Section: Unseen Action Classificationmentioning

confidence: 99%

Object Priors for Classifying and Localizing Unseen Actions

2021

View full text Add to dashboard Cite

This work strives for the classification and localization of human actions in videos, without the need for any labeled video training examples. Where existing work relies on transferring global attribute or object information from seen to unseen action videos, we seek to classify and spatio-temporally localize unseen actions in videos from image-based object information only. We propose three spatial object priors, which encode local person and object detectors along with their spatial relations. On top we introduce three semantic object priors, which extend semantic matching through word embeddings with three simple functions that tackle semantic ambiguity, object discrimination, and object naming. A video embedding combines the spatial and semantic object priors. It enables us to introduce a new video retrieval task that retrieves action tubes in video collections based on user-specified objects, spatial relations, and object size. Experimental evaluation on five action datasets shows the importance of spatial and semantic object priors for unseen actions. We find that persons and objects have preferred spatial relations that benefit unseen action localization, while using multiple languages and simple object filtering directly improves semantic matching, leading to state-of-the-art results for both unseen action classification and localization.

show abstract

“…The concept detectors are applied to the test dataset and the query is represented as a set of related concepts. This representation can be obtained using knowledge bases, such as Wikipedia [3] or EventNet [61], a semantic embedding, such as VideoStory [16] or word2vec [33], or a manual mapping [63]. Often, a linear weighted sum is used to score the videos on relevance to the query [19,63].…”

Section: Video Event Retrievalmentioning

confidence: 99%

Improving video event retrieval by user feedback

Boer

Pingen

Knook

et al. 2017

Multimed Tools Appl

Self Cite

View full text Add to dashboard Cite

In content based video retrieval videos are often indexed with semantic labels (concepts) using pre-trained classifiers. These pre-trained classifiers (concept detectors), are not perfect, and thus the labels are noisy. Additionally, the amount of pre-trained classifiers is limited. Often automatic methods cannot represent the query adequately in terms of the concepts available. This problem is also apparent in the retrieval of events, such as bike trick or birthday party. Our solution is to obtain user feedback. This user feedback can be provided on two levels: concept level and video level. We introduce the method Adaptive Relevance Feedback (ARF) on video level feedback. ARF is based on the classical Rocchio relevance feedback method from Information Retrieval. Furthermore, we explore methods on concept level feedback, such as the re-weighting and Query Point Modification (QPM) methods as well as a method that changes the semantic space the concepts are represented in. Methods on both concept level and video level are evaluated on the international benchmark TRECVID Multimedia Event Detection (MED) and compared to state of the art methods. Results show that relevance feedback on both concept and video level improves performance compared to using no relevance feedback; relevance feedback on video level obtains higher performance compared to relevance feedback on concept level; our proposed ARF method on video level outperforms a state of the art k-NN method, all methods on concept level and even manually selected concepts.

show abstract

Knowledge based query expansion in complex multimedia event detection

Cited by 22 publications

References 37 publications

Query expansion techniques for information retrieval: A survey

Query expansion techniques for information retrieval: A survey

Object Priors for Classifying and Localizing Unseen Actions

Improving video event retrieval by user feedback

Contact Info

Product

Resources

About