Building the gist of a scene: the role of global image features in recognition.

By Aude Oliva, Antonio Torralba. 2006

1,622 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Cognitive Neuroscience

In Computer Vision and Pattern Recognition

Citations in the corpus

Context Encoders: Feature Learning by Inpainting (2016, methodology): “Furthermore, previous completion relies on a hand-crafted distance metric, such as Gist [31] for nearest-neighbor computation which is inferior to a learned distance metric.”
Learning to See by Moving (2015, methodology): “The KITTI-Net outperforms hand-engineered GIST [31] features developed for scene classiﬁcation, but is outper-formed by Dense SIFT with spatial pyramid matching (SPM) kernel [23].”
Exploring Nearest Neighbor Approaches for Image Captioning (2015, methodology): “Which images are “nearest” can be deﬁned in several ways, and we examine using GIST [28], pre-trained deep features [32], and deep features ﬁne-tuned for the task of caption generation [8].”
Determinantal Point Processes for Machine Learning (2012, methodology): “• GIST: The images are processed using code from Oliva and Torralba [118] to yield 960-dimensional GIST feature vectors characterizing properties like “openness,” “roughness,” “naturalness,” and so on.”
IM2GPS: estimating geographic information from a single image (2008, background): “Gist Descriptor + Color: The gist descriptor [11] has been shown to work well for scene categorization [10] and for retrieving semantically and structurally similar scenes [4].”
An empirical study of context in object detection (2009, background): “Another type of context, the scene gist proposed by Oliva and Torralba [26], uses global edge energy statistics of an image capturing the “gist” of the visual experience [25, 29].”
Learning a Fine Vocabulary (2010, methodology): “Global methods like GIST [10] achieve a much smaller memory footprint at the cost of allowing whole image queries only.”
Efficient Object Category Recognition Using Classemes (2010, methodology): “The GIST descriptor [19] is applied to color images.”
The natural/man-made distinction is made before basic-level distinctions in scene gist processing (2010, background): “First, the current results use human subject data to test a central assumption of an influential computational model of scene gist recognition, the Spatial Envelope model (Oliva & Torralba, 2001, 2006; Oliva et al., 2003; Torralba, 2003; Torralba et al., 2006), which has been essentially untested.”
Every Picture Tells a Story: Generating Sentences from Images (2010, positive): “We encode global information of images using gist [21].”
Semantic colorization with internet images (2011, methodology): “Rather than employ the filtering method used for foreground objects, we utilize the compact GIST descriptor [Oliva and Torralba 2006], which has been shown to be effective at finding semantically similar scenes such as forests and beaches.”
Parahippocampal and retrosplenial contributions to human spatial navigation (2008, background): “PPA solely encodes geometric information, or whether it also encodes information about the distribution of visual features such as colors or textures within a scene [ 47 ].”
Data-driven visual similarity for cross-domain image matching (2011, methodology): “…Transactions on Graphics, Vol. 30, No. 6, Article 154, Publication date: December 2011. ious image feature representations (SIFT [Lowe 2004], GIST [Oliva and Torralba 2006], HoG [Dalal and Triggs 2005], wavelets, etc.) that aim to capture the locally salient (i.e., high gradient and high…”