Interactive Training And Architecture For Deep Object Selection

Forte, Marco; Price, Brian; Cohen, Scott; Xu, Ning; Pitié, François

doi:10.1109/icme46284.2020.9102942

Cited by 8 publications

(21 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…GCAMatting [26] designs a Guided Contextual Attention module to capture contextual affinity information to estimate the alpha matte of semi-transparent objects. FBAMatting [13] designs a network to estimate the foreground color, background color and alpha matte at the same time, and uses a first-order approximation to the Bayesian formula to refine the prediction results. HDMatt [45] introduces a Cross-Patch Contextual module to improve the performance under patch-based inference.…”

Section: Related Workmentioning

confidence: 99%

“…To refines the details of the upsampled feature maps, we use skip-connections to incorporate the low-level features. Finally, we stack three convolutional layers to estimate the foreground 𝑭 , background 𝑩 the alpha matte 𝜶 following FBAMatting [13].…”

Section: Matting Decodermentioning

confidence: 99%

“…Since the foreground F , background B, and alpha matte 𝛼 are unknown, only the observed image I is known, image matting is an ill-defined problem. To address this problem, most existing methods [3,5,8,12,13,15,19,25,26,28,41,44,45] need a trimap to indicate the known foreground and background regions and unknown regions in the image, which can be roughly categorized into sampling-based methods, propagation-based methods, and learning-based methods. Sampling-based methods [2,9,12,15,31,36,41] select the best foreground and background color pairs in the known regions for the each pixel in unknown regions to estimate the alpha values based on assumptions on color association.…”

Section: Introductionmentioning

confidence: 99%

“…However, all the above methods are sensitive to the foreground and background color distributions overlap that are common in natural images, which leads to unexpected artifacts in the predicted alpha mattes [44]. Recently, learning-based methods [3,8,13,26,28,44,45] use neural networks to learn both the color information and natural structure from well-annotated datasets, which achieve better performance than traditional methods. However, these methods only learn local image features due to the small effective reception fields of neural networks, which ignore the long-range image features outside the reception fields.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Long-Range Feature Propagating for Natural Image Matting

Liu

Xie

Zhang

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Natural image matting estimates the alpha values of unknown regions in the trimap. Recently, deep learning based methods propagate the alpha values from the known regions to unknown regions according to the similarity between them. However, we find that more than 50% pixels in the unknown regions cannot be correlated to pixels in known regions due to the limitation of small effective reception fields of common convolutional neural networks, which leads to inaccurate estimation when the pixels in the unknown regions cannot be inferred only with pixels in the reception fields.To solve this problem, we propose Long-Range Feature Propagating Network (LFPNet), which learns the long-range context features outside the reception fields for alpha matte estimation. Specifically, we first design the propagating module which extracts the context features from the downsampled image. Then, we present Center-Surround Pyramid Pooling (CSPP) that explicitly propagates the context features from the surrounding context image patch to the inner center image patch. Finally, we use the matting module which takes the image, trimap and context features to estimate the alpha matte. Experimental results demonstrate that the proposed method performs favorably against the state-of-the-art methods on the AlphaMatting and Adobe Image Matting datasets.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Matting Decodermentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Long-Range Feature Propagating for Natural Image Matting

Liu

Xie

Zhang

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

show abstract

“…Finally, a standard geodesic path solver (Cohen, 2006), applied on the boundaries map refines the segmentation. A recent solution (Forte et al, 2020) builds upon a U‐Net (Ronneberger et al, 2015) architecture reaches an exceptional accuracy of between 95% and 99% of mean intersection‐over‐union (mIoU), a measure of overlap between labelled regions, by using an elevated number of user's clicks (around 20). A solution between bounding boxes and point‐clicks is represented by clicking the object's extremes (top, left, bottom, and right).…”

Section: Related Workmentioning

confidence: 99%

TagLab: AI‐assisted annotation for the fast and accurate semantic segmentation of coral reef orthoimages

Pavoni

Corsini

Ponchio

et al. 2021

Journal of Field Robotics

View full text Add to dashboard Cite

Semantic segmentation is a widespread image analysis task; in some applications, it requires such high accuracy that it still has to be done manually, taking a long time. Deep learning‐based approaches can significantly reduce such times, but current automated solutions may produce results below expert standards. We propose agLab, an interactive tool for the rapid labelling and analysis of orthoimages that speeds up semantic segmentation. TagLab follows a human‐centered artificial intelligence approach that, by integrating multiple degrees of automation, empowers human capabilities. We evaluated TagLab's efficiency in annotation time and accuracy through a user study based on a highly challenging task: the semantic segmentation of coral communities in marine ecology. In the assisted labelling of corals, TagLab increased the annotation speed by approximately 90% for nonexpert annotators while preserving the labelling accuracy. Furthermore, human–machine interaction has improved the accuracy of fully automatic predictions by about 7% on average and by 14% when the model generalizes poorly. Considering the experience done through the user study, TagLab has been improved, and preliminary investigations suggest a further significant reduction in annotation times.

show abstract