Filtering Feature Mismatches using Multiple Descriptors

Similar-Part Approximation Using Invariant Feature Descriptors

This thesis describes the implementation of a new algorithm for identifying similar parts in multiple images. Visual words of Scale Invariant Feature Transform descriptors are used to describe parts of images and Locality Sensitive Hashing is utilized for quick similar-part lookup. This algorithm proves very effective on identical items of different sizes and rotations, and moderately effective...

Shape matching and retrieval based on multiple feature descriptors

A lot of 3D shape descriptors for 3D shape retrieval have been presented so far. This paper proposes a new mechanism, which employs several existing global and local 3D shape descriptors as input. With the sparse theory, some descriptors which play the most important role in measuring similarity between query model and the model in the dataset are selected automatically and an affinity matrix i...

Fish Shape Recognition using Multiple Shape Descriptors

This paper studies recognition of fish shapes using both Region based and Contour based shape based descriptors[9]. Moment Invariants are chosen as the Region based descriptor and the Simple (geometric) shape descriptors (SSD) are used as Contour based shape descriptors. The shapes are varied through scaling and rotation. Manhattan Distance is used as the classifier. The study of the recognitio...

Feature Identities, Descriptors and Handles

A Comparison of Feature Descriptors

Local Features provide a compact description of an image and are widely used in a number of vision related tasks like stereo reconstruction, objection recognition, tracking, etc. Owing to its importance, it has received considerable attention in the computer vision research community and they have been evaluated in restricted settings. In this work we do two things (1) Do a study of feature des...