Dimah V Yanovsky

Title: Dimah V. Yanovsky: Image Processing and Viewpoint-Based Compression

Introduction

Dimah V. Yanovsky is an inventor associated with Plainsboro, New Jersey, and Lucent Technologies Inc. Her patent work addresses methods for matching image features across views and compressing images captured from different viewpoints. She is credited with 3 patents.

Latest Patents

One of Yanovsky’s patents, “Method and Apparatus for Finding Feature Correspondences Between Images Captured in Real-World Environments,” describes techniques for identifying consistent feature matches across images taken from different viewpoints. The method tracks features between neighboring images, identifies potential correspondences, and uses a greedy graph-labeling algorithm to select a globally consistent set. These correspondences can then support image warping and combination to create interpolated views of an environment.

Her patent “Method and Apparatus for Compressing and Decompressing Images Captured from Viewpoints Throughout N-Dimensional Space” focuses on image compression across a multidimensional range of viewpoints. It combines a hierarchical image structure with compression methods, arranging reference images and residual images in structures such as binary trees. Images can be reconstructed through sequences of warping and addition operations, enabling access to views along different paths through the space.

Career Highlights

Yanovsky’s listed work at Lucent Technologies Inc. connects her inventions to computer graphics and image-processing applications. Her patents describe approaches intended to make image correspondence, storage, and reconstruction more efficient across varied viewpoints.

Collaborations

Yanovsky’s listed collaborators include Daniel G. Aliaga and Ingrid Birgitta Carlbom. Their shared patent-related work reflects collaboration on techniques for image processing and computer graphics.

Conclusion

Dimah V. Yanovsky’s patent portfolio highlights methods for establishing consistent image correspondences and efficiently compressing and reconstructing images across multidimensional viewpoints.