Similarity measure for image resizing using SIFT feature

Abstract

On the basis of the Scale Invariant Feature Transform (SIFT) feature, we research the distance measure in the process of image resizing. Through extracting SIFT features from the original image and the resized one, respectively, we match the SIFT features between two images, and calculate the distance for SIFT feature vectors to evaluate the degree of similarity between the original and the resized image. On the basis of the Euclidean distance measure, an effective image resizing algorithm combining Seam Carving with Scaling is proposed. We first resize an image using Seam Carving, and calculate the similarity distance between the original image and its resized one. Before the salient object and content are damaged obviously, we stop Seam Carving and transfer residual task to Scaling. Experiments show that our algorithm is able to avoid the damage and distortion of image content and preserve both the local structure and the global visual effect of the image graciously.

Automatic topics segmentation for TV news video using prior knowledge

“…A segment localized at frame j is identified as program type t only if the frames signatures of Vsig(t) are similar to their homologues in the stream as detailed in similarity measurement (11).…”

“…For SigF POI signature similarity measurement (5), we use an average distance DistSIFT (FS,GS) [11] comparing all SIFT features from a frame stream (FS) and their homologues extracted from video grammar (GS). This measurement is suitable in cases of number variation of detected feature points of resized images.…”

3D Reconstructions Using Unstabilized Video Footage from an Unmanned Aerial Vehicle

“…During initial investigations by the authors, such a low frame rate was also deemed insufficient to obtain the necessary overlap in unstabilized footage. Although this video footage has a lower resolution than still images, decreasing the image size does not necessarily degrade the performance of the SIFT Algorithm [21], because VisualSFM itself will re-size any image above 3200 pixels before processing [6]. Notably, the GoPro camera has a fixed wide angle lens, which causes significant distortion of the image on the CCD.…”

Image resizing with SIFT feature preservation

“…Hua et al [8] computed the Euclidean distances of SIFT feature vectors between an original image and the corresponding resized image for evaluation of deformation of the resized image. They use the distance as the criterion for combining scaling and seam carving approaches.…”