Distinctive Image Features from Scale-Invariant Keypoints

By David G. Lowe. 2004

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

Aggregated Residual Transformations for Deep Neural Networks (2016, background): “In contrast to traditional hand-designed features ( e.g ., SIFT [29] and HOG [5]), features learned by neural networks from large-scale data [33] require minimal human involvement during training, and can be transferred to a variety of recognition tasks [7, 10, 28].”
Understanding deep image representations by inverting them (2014, background): “Most image understanding and computer vision methods build on image representations such as textons [15], histogram of oriented gradients (SIFT [18] and HOG [4]), bag of visual words [3][25], sparse [35] and local coding [32], super vector coding [37], VLAD [9], Fisher Vectors [21], and, lately,…”
A Multi-View Embedding Space for Modeling Internet Images, Tags, and Their Semantics (2012, methodology): “We will refer to these six features as D-SIFT, D-CSIFT, D-RGBSIFT, H-SIFT, HCSIFT, and H-RGBSIFT.”
A self-organizing neural network architecture for learning human-object interactions (2017, positive): “We extract dense SIFT features, which are not more than SIFT descriptors [38] computed at crossing points of ﬁxed grids superimposed on each object image 1 .”
Very Deep Convolutional Networks for Text Classification (2016, methodology): “In early approaches to computer vision, handcrafted features were used, for instance “scale-invariant feature transform (SIFT)”(Lowe, 2004), followed by some classifier.”
The AdobeIndoorNav Dataset: Towards Deep Reinforcement Learning based Real-world Indoor Robot Visual Navigation (2018, methodology): “We also detect SIFT key points [29] from all visual inputs to identify targets with distinctive image features.”
Deep roto-translation scattering for object classification (2014, methodology): “Learning image representations has considerably enhanced image classiﬁcation results compared to geometric features such as edge descriptors, or, SIFT and HOG [17, 5] patch representations.”
Meta-Learning in Neural Networks: A Survey (2020, background): “Historically, the success ofmachine learningwas driven by advancing hand-engineered features [11], [12].”
Cut, Paste and Learn: Surprisingly Easy Synthesis for Instance Detection (2017, methodology): “Early approaches, such as [6], heavily depend on extracting local features such as SIFT [30], SURF [3], MSER [32] and matching them to retrieve instances [29, 48].”
Polar Transformer Networks (2017, background): “It starts with the basic idea inherent in SIFT [18] about equivariance of anchor points such as the maxima of filtered responses in (translation) space.”
Surpassing Human-Level Face Verification Performance on LFW with GaussianFace (2014, methodology): “For the methods in the ﬁrst category, for example, low-level features such as SIFT [36], LBP [3], and Gabor [34] are handcrafted.”
Situated Dialogue and Spatial Organization: What, Where… and Why? (2007, methodology): “The vision subsystem uses an implementation of SIFT (Scale Invariant Feature Transform) features (Lowe, 2004) for recognizing typical objects in the environment like television sets, coffee machines or bookcases.”
A performance evaluation of gradient field HOG descriptor for sketch based image retrieval (2013, methodology): “These are matched in a nearest-neighbor manner within the descriptor space via Lowe’s SIFT matching scheme[62] i.”
TAPIR: Tracking Any Point with per-frame Initialization and temporal Refinement (2023, methodology): “Keypoint Correspondence methods aim at sparse keypoint matching given image pairs, from hand-defined discriminative feature descriptors [3, 32, 33, 49] to deep learning based methods [11, 25, 35, 40].”
LoFTR: Detector-Free Local Feature Matching with Transformers (2021, background): “SIFT [26] and ORB [35] are arguably the most successful hand-crafted local features and are widely adopted in many 3D computer vision tasks.”