Applications of Scene Attributes
نویسنده
چکیده
In this paper, we study the feasibility of scene attributes as the intermediate scene representation for automatic image captioning, tag predicting and semantic image search. we show that when used as features for these tasks, low dimensional scene attributes can compete with or improve on the state of art performance. In particular, we propose a new method of content-based image retrieval, which takes advantage of the correlation between scene attributes and caption key words. When compared to simple uni-gram tfidf image search, our method offers more promising results.
منابع مشابه
Compressed Domain Scene Change Detection Based on Transform Units Distribution in High Efficiency Video Coding Standard
Scene change detection plays an important role in a number of video applications, including video indexing, searching, browsing, semantic features extraction, and, in general, pre-processing and post-processing operations. Several scene change detection methods have been proposed in different coding standards. Most of them use fixed thresholds for the similarity metrics to determine if there wa...
متن کاملDerived fuzzy importance of attributes based on the weakest triangular norm-based fuzzy arithmetic and applications to the hotel services
The correlation between the performance of attributes and the overallsatisfaction such as they are perceived by the customers is often used tocalculate the importance of attributes in the crisp case. Recently, the methodwas extended, based on the standard Zadeh extension principle, to the fuzzycase, taking into account the specificity of the human thinking. Thedifficulties of calculation are im...
متن کاملHierarchical Space Tiling (HST) for scene classification and attribute localization
In this paper, we propose a representation named Hierarchical Space Tiling (HST) and a weakly supervised method for simultaneously leaning the scene configurations and attributes from a collection of natural images associated with attributes in text, where the precise localization of each attributes left unknown. We start with an over-complete HST which quantizes the huge and continuous scene c...
متن کاملBuilding a Taxonomy of Attributes for Fine-Grained Scene Understanding
This paper presents the first effort to discover and exploit a diverse taxonomy of scene attributes. Starting with the fine-grained SUN database, we perform crowd-sourced human studies to find over 100 attributes that discriminate between scene categories. We construct an attributelabeled dataset on top of the SUN database [7]. This “SUN Attribute database” spans more than 700 categories and 14...
متن کاملTransient Attributes or High-Level Understanding and Editing of Outdoor Scenes
We live in a dynamic visual world where the appearance of scenes changes dramatically from hour to hour or season to season. In this work we study “transient scene attributes” – high level properties which affect scene appearance, such as “snow”, “autumn”, “dusk”, “fog”. We define 40 transient attributes and use crowdsourcing to annotate thousands of images from 101 webcams. We use this “transi...
متن کامل