Halfway through the semantic gap: Prosemantic features for image retrieval

Ciocca, Gianluigi; Cusano, Claudio; Santini, Simone; Schettini, Raimondo

doi:10.1016/j.ins.2011.06.025

Cited by 28 publications

(7 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One possibility, which we shall not analyze in this paper, is that of a prosemantic space, in which each dimension in the reduced dimensionality space is the output of a classifier, trained to recognize a specific category of images [8,9]. Here, we shall consider the individual feature spaces as given, and use the query to transform each one into a dimension of the semantic space.…”

Section: Semantic Partitionmentioning

confidence: 99%

Relevance Feedback with Latent Variables in Riemann spaces

Santini

2019

Preprint

Self Cite

View full text Add to dashboard Cite

In this paper we develop and evaluate two methods for relevance feedback based on endowing a suitable "semantic query space" with a Riemann metric derived from the probability distribution of the positive samples of the feedback. The first method uses a Gaussian distribution to model the data, while the second uses a more complex Latent Semantic variable model. A mixed (discrete-continuous) version of the Expectation-Maximization algorithm is developed for this model.We motivate the need for the semantic query space by analyzing in some depth three well-known relevance feedback methods, and we develop a new experimental methodology to evaluate these methods and compare their performance in a neutral way, that is, without making assumptions on the system in which they will be embedded.

show abstract

Section: Semantic Partitionmentioning

confidence: 99%

Relevance Feedback with Latent Variables in Riemann spaces

Santini

2019

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…These attribute detectors are then run on new images for high level recognition [46,47]. Researchers have explored creating a set (or bank) of detectors pretrained on objects such as Object Banks [27], an ontology of abstract concepts such as Classemes [45] or scene attributes [8,37].…”

Section: Related Workmentioning

confidence: 99%

Complex Event Recognition from Images with Few Training Examples

Ahsan

Sun

Hays

et al. 2017

2017 IEEE Winter Conference on Applications of Computer Vision (WACV)

View full text Add to dashboard Cite

We propose to leverage concept-level representations for complex event recognition in photographs given limited training examples. We introduce a novel framework to discover event concept attributes from the web and use that to extract semantic features from images and classify them into social event categories with few training examples. Discovered concepts include a variety of objects, scenes, actions and event sub-types, leading to a discriminative and compact representation for event images. Web images are obtained for each discovered event concept and we use (pretrained) CNN features to train concept classifiers. Extensive experiments on challenging event datasets demonstrate that our proposed method outperforms several baselines using deep CNN features directly in classifying images into events with limited training examples. We also demonstrate that our method achieves the best overall accuracy on a dataset with unseen event categories using a single training example.

show abstract

“…Finally, the recognize scores are packed together as prosemantic features for indexing into image retrieval system. [4] When user wants to search for semantically similar images, it might be difficult using only low-level image features. For example, user wants to query "bear on a river".…”

Section: Introductionmentioning

confidence: 99%

Image Annotation Fusing Content-Based and Tag-Based Technique Using Support Vector Machine and Vector Space Model

Chan

Yamana

et al. 2014

2014 Tenth International Conference on Signal-Image Technology and Internet-Based Systems

View full text Add to dashboard Cite

In this paper, we propose a new image annotation method by combining content-based image annotation and tag-based image annotation techniques. Content-based image annotation technique is adopted to extract "loosely defined concepts" by analyzing pre-given images' features such as color moment (CM), edge orientation histogram (EOH), and local binary pattern (LBP); followed by constructing a set of SVMs for 100 loosely defined concepts. A base-vector for each concept, similar to tag-based image annotation technique, is then constructed by using SVMs' predicted probabilistic results for sample-images whose main concepts are known. Finally cosine similarity between a query-image vector and the base vector is calculated for each concept. Experimental results show that our proposed method outperforms content-based image annotation technique by about 23% in accuracy.

show abstract

Halfway through the semantic gap: Prosemantic features for image retrieval

Cited by 28 publications

References 41 publications

Relevance Feedback with Latent Variables in Riemann spaces

Relevance Feedback with Latent Variables in Riemann spaces

Complex Event Recognition from Images with Few Training Examples

Image Annotation Fusing Content-Based and Tag-Based Technique Using Support Vector Machine and Vector Space Model

Contact Info

Product

Resources

About