GIF image retrieval in cloud computing environment

Abstract : GIF images have been used in the last years, especially on social media. Here it is explored a content-based image retrieval system to work specifically with GIF file format. Its implementation is extended to a cloud computing environment. Given the Tumblr GIF dataset, it is created a "search by example" image retrieval system. To describe the images, low-level features are used: (1) color, (2) texture and (3) shape. The system performs the search using just GIF images as query images. To obtain faster results on the retrieval process, a hashing indexing approach is used. The system showed a complexity of O(n 2) for indexing and O(log(n)) for retrieval. Additionally, better results were obtained (in relation to precision and recall) for simple images, instead of images with a lot of movements.
Complete list of metadatas

Cited literature [17 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02349629
Contributor : Evelyn Paiz <>
Submitted on : Tuesday, November 5, 2019 - 5:19:58 PM
Last modification on : Friday, November 8, 2019 - 3:06:24 PM

File

paper345.pdf
Files produced by the author(s)

Identifiers

Citation

Evelyn Paiz-Reyes, Nadile Nunes-De-Lima, Sule Yildirim-Yayilgan. GIF image retrieval in cloud computing environment. Image Analysis and Recognition, pp.261-268, 2018, ⟨10.1007/978-3-319-93000-8_30⟩. ⟨hal-02349629⟩

Share

Metrics

Record views

8

Files downloads

17