This paper presents a method for extracting distinctive invariant features from images that can be used for reliable matching between different views of an object or scene. The features are invariant to image scale and rotation, and are robust to affine distortion, changes in 3D viewpoint, noise, and illumination. They are highly distinctive, allowing a single feature to be correctly matched with high probability against a large database of features. The paper also describes an approach to using these features for object recognition, which involves matching individual features to a database of known objects using a fast nearest-neighbor algorithm, followed by a Hough transform to identify clusters belonging to a single object, and finally performing verification through least-squares solution for consistent pose parameters. This approach allows robust identification of objects among clutter and occlusion while achieving near real-time performance. The method involves four stages: scale-space extrema detection, keypoint localization, orientation assignment, and keypoint descriptor. Scale-space extrema detection identifies potential interest points invariant to scale and orientation. Keypoint localization determines the location and scale of each keypoint. Orientation assignment assigns one or more orientations to each keypoint based on local image gradient directions. The keypoint descriptor is a representation of local image gradients that allows for significant levels of local shape distortion and change in illumination. The method is named the Scale Invariant Feature Transform (SIFT) as it transforms image data into scale-invariant coordinates relative to local features. The approach generates a large number of features that densely cover the image over the full range of scales and locations. A typical image of size 500x500 pixels will give rise to about 2000 stable features. These features are highly distinctive, allowing a single feature to be correctly matched with high probability against a large database of features, providing a basis for object and scene recognition. The method is efficient and robust, with a cascade filtering approach that minimizes the cost of extracting these features. The keypoint descriptors are highly distinctive, allowing a single feature to find its correct match with good probability in a large database of features. However, in a cluttered image, many features from the background will not have any correct match in the database, giving rise to many false matches in addition to the correct ones. The correct matches can be filtered from the full set of matches by identifying subsets of keypoints that agree on the object and its location, scale, and orientation in the new image. The probability that several features will agree on these parameters by chance is much lower than the probability that any individual feature match will be in error. The determination of these consistent clusters can be performed rapidly by using an efficient hash table implementation of the generalized Hough transform. Each cluster of 3 or more features that agree on an object and its pose is then subject to further detailed verification. First, a least-squares estimate is made for an affine approximation to the object pose. Any other image features consistent with this pose are identified, and outliers areThis paper presents a method for extracting distinctive invariant features from images that can be used for reliable matching between different views of an object or scene. The features are invariant to image scale and rotation, and are robust to affine distortion, changes in 3D viewpoint, noise, and illumination. They are highly distinctive, allowing a single feature to be correctly matched with high probability against a large database of features. The paper also describes an approach to using these features for object recognition, which involves matching individual features to a database of known objects using a fast nearest-neighbor algorithm, followed by a Hough transform to identify clusters belonging to a single object, and finally performing verification through least-squares solution for consistent pose parameters. This approach allows robust identification of objects among clutter and occlusion while achieving near real-time performance. The method involves four stages: scale-space extrema detection, keypoint localization, orientation assignment, and keypoint descriptor. Scale-space extrema detection identifies potential interest points invariant to scale and orientation. Keypoint localization determines the location and scale of each keypoint. Orientation assignment assigns one or more orientations to each keypoint based on local image gradient directions. The keypoint descriptor is a representation of local image gradients that allows for significant levels of local shape distortion and change in illumination. The method is named the Scale Invariant Feature Transform (SIFT) as it transforms image data into scale-invariant coordinates relative to local features. The approach generates a large number of features that densely cover the image over the full range of scales and locations. A typical image of size 500x500 pixels will give rise to about 2000 stable features. These features are highly distinctive, allowing a single feature to be correctly matched with high probability against a large database of features, providing a basis for object and scene recognition. The method is efficient and robust, with a cascade filtering approach that minimizes the cost of extracting these features. The keypoint descriptors are highly distinctive, allowing a single feature to find its correct match with good probability in a large database of features. However, in a cluttered image, many features from the background will not have any correct match in the database, giving rise to many false matches in addition to the correct ones. The correct matches can be filtered from the full set of matches by identifying subsets of keypoints that agree on the object and its location, scale, and orientation in the new image. The probability that several features will agree on these parameters by chance is much lower than the probability that any individual feature match will be in error. The determination of these consistent clusters can be performed rapidly by using an efficient hash table implementation of the generalized Hough transform. Each cluster of 3 or more features that agree on an object and its pose is then subject to further detailed verification. First, a least-squares estimate is made for an affine approximation to the object pose. Any other image features consistent with this pose are identified, and outliers are