OpenVL is the future of developer-friendly computer vision - existing vision frameworks provide access at a very low level, such as individual algorithm names (often named after their inventor), while OpenVL provides a higher-level abstraction to hide the details of sophisticated vision techniques: developers use a task-centred API to supply a description of the problem, and OpenVL interprets the description and provides a solution.

The OpenVL computer vision abstraction will support hardware acceleration and multiple platforms (mobile, cloud, desktop, console), and therefore also allows vendor-specific implementations. We are committed to making it an open API available to everyone (and hope to make it an open standard); Continue reading...
Face Front
Face Profile
Classification of Image Registration Problems Using Support Vector Machines
Steve Oldridge, Sidney Fels and Gregor Miller
This paper introduces a system that automatically classifies image pairs based on the type of registration required to align them. The system uses support vector machines to classify between panoramas, high-dynamic-range images, focal stacks, super-resolution, and unrelated image pairs. A feature vector was developed to describe the images, and 1100 pairs were used to train and test the system with 5-fold cross validation. The system is able to classify the desired registration application using a 1:Many classifier with an accuracy of 91.18%. Similarly 1:1 classifiers were developed for each class with classification rates as follows: Panorama image pairs are classified at 93.15%, high-dynamic-range pairs at 97.56%, focal stack pairs at 95.68%, super-resolution pairs at 99.25%, and finally unrelated image pairs at 95.79%. An investigation into feature importance outlines the utility of each feature individually. In addition, the invariance of the classification system towards the size of the image used to calculate the feature vector was explored. The classification of our system remains level at approximately 91% until the image size is scaled to 10% (150 x 100 pixels), suggesting that our feature vector is image size invariant within this range.

Presented in Keauhou, Kailua-Kona, January 2011 at the Workshop on the Applications of Computer Vision.
    author = {Steve Oldridge and Sidney Fels and Gregor Miller},
    title = {Classification of Image Registration Problems Using Support Vector Machines},
    booktitle = {Proceedings of the 11th Workshop on the Applications of Computer Vision (WACV)},
    series = {WVM'11},
    pages = {360--366},
    month = {January},
    year = {2011},
    publisher = {IEEE},
    address = {New York City, New York, U.S.A.},
    isbn = {978-1-4244-9496-5},
    location = {Keauhou, Kailua-Kona, Hawai'i, U.S.A.},
    doi = {},
    url = {}