Abstract
Density-based clustering algorithms have been the most commonly used algorithms for discovering regions and points of interest in cities using global positioning system (GPS) information in geo-tagged photos. However, users sometimes find more specific areas of interest using real objects captured in pictures. Recent advances in deep learning technology make it possible to recognize these objects in photos. However, since deep learning detection is a very time-consuming task, simply combining deep learning detection with density-based clustering is very costly. In this paper, we propose a novel algorithm supporting deep content and density-based clustering, called deep density-based spatial clustering of applications with noise (DeepDBSCAN). DeepDBSCAN incorporates object detection by deep learning into the density clustering algorithm using the nearest neighbor graph technique. Additionally, this supports a graph-based reduction algorithm that reduces the number of deep detections. We performed experiments with pictures shared by users on Flickr and compared the performance of multiple algorithms to demonstrate the excellence of the proposed algorithm.
Highlights
With recent advances in mobile devices, sharing geo-tagged photos has become popular on social network services such as Facebook, Twitter, and Instagram
We present deep content and density-based clustering techniques for geo-tagged photos
Since spatial image data sets are stored in large quantities in systems such as databases, the time taken to discover a particular class of regions of interest is added to the existing density-based spatial clustering of applications with noise (DBSCAN) techniques, as is the time spent on image detection tasks
Summary
With recent advances in mobile devices, sharing geo-tagged photos has become popular on social network services such as Facebook, Twitter, and Instagram. The following SQL code shows one of the most commonly used queries for density-based clustering of tour photos using a spatial predicate function ST_within() This query selects only photos that were taken within Yellowstone National Park. Integrating deep object detection with clustering queries could provide researchers with very powerful and convenient features in many applications When they want to find only photos containing bears within a specific region, they can make this query by adding or changing the parameters in the deep and spatial predicate functions. The traditional naive approach first performs deep content detection for all photos to construct a data set that has information about all objects detected in the photos in the batch preprocessing step This data set can be inserted into a table in the database system or searched itself when the deep predicate is evaluated in the algorithm.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
