Error-tolerant interactive image segmentation using dynamic and iterated graph-cuts

Şener, Ozan; Uǧur, Kemal; Alatan, A. Aydın

doi:10.1145/2390821.2390826

Cited by 20 publications

(22 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…APP is the similarity measure proposed An and Pellacini [AP08] used with the same dense CRF as our approach. SEN is the method of Setter et al [ŞUA12]…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Accurate Binary Image Selection from Inaccurate User Input

Subr

Paris

Soler

et al. 2013

Computer Graphics Forum

View full text Add to dashboard Cite

(a) input image + scribbles (b) our result (c) using diffusion maps (d) using KNN mattingWe present a technique to make binary selections in images, such as to select the three penguins, using inaccurate scribbles to indicate foreground (blue) and background (red). Unlike existing approaches, our approach does not assume that the indications are 100% accurate. Since the related work, diffusion maps [FFL10] as well as KNN matting [CLT12], produce fuzzy selections, we manually thresholded their results to achieve the best possible selections. AbstractSelections are central to image editing, e.g., they are the starting point of common operations such as copy-pasting and local edits. Creating them by hand is particularly tedious and scribble-based techniques have been introduced to assist the process. By interpolating a few strokes specified by users, these methods generate precise selections. However, most of the algorithms assume a 100% accurate input, and even small inaccuracies in the scribbles often degrade the selection quality, which imposes an additional burden on users. In this paper, we propose a selection technique tolerant to input inaccuracies. We use a dense conditional random field (CRF) to robustly infer a selection from possibly inaccurate input. Further, we show that patch-based pixel similarity functions yield more precise selection than simple point-wise metrics. However, efficiently solving a dense CRF is only possible in low-dimensional Euclidean spaces, and the metrics that we use are high-dimensional and often non-Euclidean. We address this challenge by embedding pixels in a low-dimensional Euclidean space with a metric that approximates the desired similarity function. The results show that our approach performs better than previous techniques and that two options are sufficient to cover a variety of images depending on whether the objects are textured.

show abstract

“…APP is the similarity measure proposed An and Pellacini [AP08] used with the same dense CRF as our approach. SEN is the method of Setter et al [ŞUA12]…”

Section: Resultsmentioning

confidence: 99%

“…Recent work by Şener et al [ŞUA12] proposes an error correction procedure and combines this with a dynamic and iterative graph‐cut algorithm for interactive segmentation. Superpixels that do not conform to a dynamically‐learned single Gaussian color model are removed from the graph.…”

Section: Introductionmentioning

confidence: 99%

Accurate Binary Image Selection from Inaccurate User Input

Subr

Paris

Soler

et al. 2013

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

“…However, we augment the interaction with the user by superimposing over the processed image both the segmentation mask generated by GrabCut and the set of SLIC [21] superpixels computed over the image itself. The user can "tap" those superpixels to decide whether they should be added or removed from the initial segmentation mask; this is similar but opposite to [16], where the interactive segmentation process directly starts from superpixels.…”

Section: A Related Workmentioning

confidence: 99%

“…While recent advancements in this field have pushed the accuracies of fully automatic techniques for solving this task much further than they were a few years ago [2], [3], in professional image editing it is still not possible to adopt fully automatic algorithms to select object boundaries as their current state-of-the-art results do not reach the minimum standard of accuracy demanded by professional image editors. For this reason, several interactive segmentation algorithms, mostly based on graph cut segmentation techniques, have been proposed over the last decade [4]- [16]. Among those, some can be guided by the users throughout the segmentation process in order to correct their errors, which often appear when the contrast between the object that needs to be selected and the background is extremely low or when the boundary of the object is highly jagged, with the aim of reaching final segmentation results that meet the desired degree of accuracy.…”

Section: Introductionmentioning

confidence: 99%

“…It is particularly interesting to observe that, even though the first published interactive segmentation methods focused on providing the highest degree of accuracy while reducing as much as possible the amount of interaction required to obtain satisfying final segmentation results [15], more recent works focused on increasing the feedback with the final users by constantly showing in real time how user inputs affect the segmentation masks throughout the segmentation process [12], [16]. In fact, real time feedback is extremely important in the image and video editing fields as the user can change the way he/she interacts with the segmentation algorithm to obtain the best results with the least effort.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Interactive Object Class Segmentation for Mobile Devices

Gallo

Zamberletti

Noce

2014

2014 27th SIBGRAPI Conference on Graphics, Patterns and Images

View full text Add to dashboard Cite

Fig. 1. The proposed method enables the user to interactively correct the errors committed by an initial GrabCut segmentation phase by "tapping" the areas of the processed image that were wrongly labeled during the initial phase; the whole method is specifically designed to produce high quality segmentations on mobile devices.Abstract-In this paper we propose an interactive approach for object class segmentation of natural images on touch-screen capable mobile devices. The key research question to which this paper tries to give an answer is: can we effectively correct the errors committed by an automatic or semi-automatic figureground segmentation algorithm while also providing real time feedback to the user on a low computational power mobile device? Many research works focused on improving automatic or semi-automatic figure-ground segmentation algorithms, but none tried to take advantage of the existing touch-screen technology integrated in most modern mobile devices to optimize the segmentation results of these algorithms. Our key idea is to use superpixels as interactive buttons that can be quickly tapped by the user to be added or removed from an initial low quality segmentation mask, with the aim of correcting the segmentation errors and produce a satisfying final result. We performed an extensive analysis of the proposed approach by implementing it both on a desktop computer and a mid-range Android device; even though our method is extremely simple, the results we obtained are comparable with those achieved by other state-of-the-art interactive segmentation algorithms. As such, we believe that the proposed approach can be exploited by most image editing mobile applications to provide a simple but highly effective method for interactive object class segmentation.

show abstract

Color Transfer

Liu

2023

Synthesis Lectures on Visual Computing: Computer Graphics, Animation, Computational Photography and Imaging

View full text Add to dashboard Cite

Error-tolerant interactive image segmentation using dynamic and iterated graph-cuts

Cited by 20 publications

References 16 publications

Accurate Binary Image Selection from Inaccurate User Input

Accurate Binary Image Selection from Inaccurate User Input

Interactive Object Class Segmentation for Mobile Devices

Color Transfer

Contact Info

Product

Resources

About