TextonBoost: Joint Appearance, Shape and Context Modeling for Multi-class Object Recognition and Segmentation

By Jamie Shotton, John M. Winn, Carsten Rother, Antonio Criminisi. 2006

1,393 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

AI and the Everything in the Whole Wide World Benchmark (2021, background): “At the time of its creation, it offered 20 times the number of categories, and 100 times the number of total images as the most popular training and evaluation benchmarks at the time, such as Caltech 101/256 [Kamarudin et al., 2015, Grifﬁn et al., 2007], MSRC [Shotton et al., 2006] and PASCAL VOC [E”
Relation Networks for Object Detection (2017, background): “(NMS), is then applied to remove duplicate detections. It has been well recognized in the vision community for years that contextual information, or relation between objects, helps object recognition [14,19,53,54,46,41,19, 18,7]. Most such works are before the prevalence of deep learning. During the”
Panoptic Segmentation (2018, background): “Semantic segmentation datasets have a rich history [35, 23, 9] and helped drive key innovations (e.”
Simultaneous Detection and Segmentation (2014, methodology): “The MSRC dataset [30] was one of the first publicly available benchmarks geared towards this task.”
COCO-Stuff: Thing and Stuff Classes in Context (2016, methodology): “In particular, MSRC 21 [46], KITTI [23], CamVid [7], SIFT Flow [36] and NYUD [47] all have less than 5,000 images (Table 1).”
Segment Anything (2023, background): “Segmentation is a broad ﬁeld: there’s interactive segmentation [57, 109], edge detection [3], super pixelization [85], object proposal generation [2], foreground segmentation [94], semantic segmentation [90], instance segmentation [66], panoptic segmentation [59], etc .”
Cross-Stitch Networks for Multi-task Learning (2016, methodology): “lti-task learning, we choose varied tasks on multiple datasets. In particular, we select four well established and diverse tasks on different types of image datasets: 1) We pair semantic segmentation [27,45,46] and surface normal estimation [11,18,52], both of which require predictions over all pixe”
YouTube-8M: A Large-Scale Video Classification Benchmark (2016, methodology): “Starting from a number of well labeled small-scale datasets such as Caltech 101/256 [8, 10], MSRC [32], PASCAL [7], image understanding research has rapidly advanced to utilizing larger datasets such as ImageNet [6] and SUN [38] for the next generation of vision algorithms.”
Unsupervised Discovery of Mid-Level Discriminative Patches (2012, methodology): “Others have employed discriminative part detectors such as poselets [17], at-Unsupervised tribute detectors [18], “visual phrases” [19], or “stuﬀ” detectors [20] as features.”
SUN Database: Exploring a Large Collection of Scene Categories (2014, background): “Texton: A traditional and powerful local image descriptor is to convolve the image with Gabor-like filter bank [41].”
Data-Driven Grasp Synthesis—A Survey (2013, background): “These can be, e.g., a common texture [62] or shape [63, 64], the occurrence of speciﬁc local features [65, 66] or their speciﬁc spatial constellation [67, 68].”
Discrete optimization of ray potentials for semantic 3D reconstruction (2015, methodology): “The semantic classiﬁer [15] for 5 classes (building, tree, ground, clutter and sky) was trained on the CamVid [4] and MSRC datasets [22], with additional training data from [11].”
MAP Estimation of Semi-Metric MRFs via Hierarchical Graph Cuts (2009, background): “related areas can be expressed using semi-metric labeling, from low level tasks like image denoising and stereo reconstruction [30] to high level tasks like pose estimation [9] and scene segmentation [27]. Hence, the semi-metric labeling problem merits special attention. We describe a novel algorith”
From contours to regions: An empirical evaluation (2009, methodology): “Region benchmarks on MSRC and PASCAL08.”
Layered object detection for multi-class segmentation (2010, methodology): “Candidate bounding boxes are often generated using a scanning window approach and scored using a classifier trained on positive and negative examples[24, 2, 4]. In contrast, multi-class segmentation models have largely been built on top of Markov Random Field (MRF) models which enforce smoothness ac”
A Hybrid Approach to Detect and Localize Texts in Natural Scene Images (2011, background): “Any two components whose spatial relationship obeys this rule are linked as neighbors in the graph.”
Relating Things and Stuff via ObjectProperty Interactions (2014, methodology): “As a result, their method works well on datasets which contains only few object instances (typically less than 3) such as the MSRC dataset [27].”