James Lee Hafner

Query by image and video content: the QBIC system

Research on ways to extend and improve query methods for image databases is widespread. We have developed the QBIC (Query by Image Content) system to explore content-based retrieval methods. QBIC allows queries on large image and video databases based on example images, user-constructed sketches and drawings, selected color and texture patterns, camera and object motion, and other graphical information. Two key properties of QBIC are (1) its use of image and video content-computable properties of color, texture, shape and motion of images, videos and their objects-in the queries, and (2) its graphical query language, in which queries are posed by drawing, selecting and other graphical means. This article describes the QBIC system and demonstrates its query capabilities. QBIC technology is part of several IBM products. >

Efficient color histogram indexing for quadratic form distance functions

An improved shipping container having novel locking features in the end panel and corner flaps. The novel locking features improved bulge resistance at the end panels from sideward bulge of the product. The improved container comprises a pair of corner flaps being hinged from the side panels and folded inwardly against an end panel with the two corner flaps and end panel on each side of the container having a quadruple lock. The lock is formed by providing a locking tab on each corner flap as well as a pair of locking tabs on the end panel with the end panel and corner flaps locking tabs being designed to be swung and locked in the opening formed by the aligned mating locking tab.In image retrieval based on color, the weighted distance between color histograms of two images, represented as a quadratic form, may be defined as a match measure. However, this distance measure i...

HoVer Erasure Codes For Disk Arrays

We present a new family of XOR-based erasure codes primarily targeted for use in disk arrays. These codes have a unique data/parity layout, with both horizontal and vertical parity arrangements giving rise to the name HoVer codes. We give constructions that tolerate up to four disk failures. Though the codes are only approximately maximum distance separable (MDS), they have performance advantages over other codes at many common array sizes. In addition, they have fewer parameter constraints than many other codes which enable greater choices and flexibility in efficiency and performance trade-offs

Asymptotically fast triangularization of matrices over rings

This paper considers the problems of triangularizing and diagonalizing matrices over rings, with particular emphasis on the integral case. It begins with a description of fast algorithms for the computation of Hermite and Smith normal forms of integer matrices. Then it shows how to apply fast matrix multiplication techniques to the problem of triangularizing a matrix over a ring using elementary column operations. These general results lead to an algorithm for triangularizing integer matrices that has a faster running time than the known Hermite normal form algorithms. The triangular matrix that is computed has small entries like the Hermite normal form, and will suffice for many applications.

Performing color space conversions with three-dimensional linear interpolation

Three-dimensionalinterpolation is suitable for many kinds of color space transformations. We examine and analyze several linear interpolation schemes-some standard, some known, and one novel. An interpolation algorithm design is divided into three parts: packing (filling the space of the input variable with sample points), extraction (selecting from the constellation of sample points those appropriate to the interpolation of a specific input point), and calculation (using the extracted values and the input point to determine the interpolated approximation to the outputpoint). We focus on regular (periodic) packing schemes. Seven principles govern the design of linear interpolation algorithms: 1) Each sample point should be used as a vertex of as many polyhedra as possible; 2) the polyhedra should completely fill the space; 3) polyhedra that share any part of a face must share the entire face; 4) the polyhedra used should have the fewest vertices possible; 5) polyhedra should be small; 6) in the absence of information about cuivature anisotropy, polyhedra should be close to regular in shape; and 7) polyhedra should be of similar size. A test for interpolation algorithm performance in performing actual color space conversions is described, and results are given for an example color space conversion using several linear interpolation methods. The extractions from cubic, body-centered-cubic, and face-centered-cubic lattices are described and analyzed. The results confirm Kanamoris claims for the accuracy of PRISM interpolation; it comes close to the accuracy of trilinear interpolation with roughly three-quarters the computations. The results show that tetrahedral interpolation, with close to half the computational cost of tnlinear interpolation, is capable of providing better accuracy. Of the tetrahedral interpolation techniques, one diagonal extraction from cubic packing is useful as a general-purpose color space interpolator...

Efficient color histogram indexing

In image retrieval based on color, the weighted distance between color histograms of two images, represented as a quadratic form, may be defined as a match measure. However, this distance measure is computationally expensive (naively O(N/sup 2/) and at best O(N) in the number N of histogram bins) and it operates on high dimensional features (O(N)). We propose the use of low-dimensional, simple to compute distance measures between the color distributions, and show that these are lower bounds on the histogram distance measure. Results on color histogram matching in large image databases show that pre-filtering with the simpler distance measures leads to significantly less time complexity because the quadratic histogram distance is now computed on a smaller set of images. The low-dimensional distance measure can also be used for indexing into the database.<<ETX>>

The query by image content (QBIC) system

QBIC (Query By Image Content) is a prototype software system for image retrieval developed at the IBM Almaden Research Center. It allows a user to query an image collection using features of image content – colors, textures, shapes, locations, and layout of images and image objects. For example, a user can query for images with a green background that contain a round red object in the upper left. The queries are formed graphically a query for red objects can be specified by selecting the color red from a color wheel, a texture query can be specified by selecting from a palette of textures, a query for a shape can be specified by drawing the shape on a” blackboard”, and so on. Retrievals are based on similarity, not exact match, computed from nuHarry Road, San Jose, CA 95120

Updates to the QBIC system

QBICTM (Query By Image Content) is a set of technologies and associated software that allows a user to search, browse, and retrieve image, graphic, and video data from large on-line collections. This paper discusses current research directions of the QBIC project such as indexing for high-dimensional multimedia data, retrieval of gray level images, and storyboard generation suitable for video. It describes aspects of QBIC software including scripting tools, application interfaces, and available GUIs, and gives examples of applications and demonstration systems using it.

Partial-MDS Codes and Their Application to RAID Type of Architectures

A family of codes with a natural 2-D structure is presented, inspired by an application of redundant arrays of independent disks (RAID) type of architectures whose units are solid-state drives (SSDs). Arrays of SSDs behave differently from arrays of hard disk drives, since hard errors in sectors are common and traditional RAID approaches (like RAID 5 or RAID 6) may be either insufficient or excessive. An efficient solution to this problem is given by the new codes presented, called partial maximum distance separable (PMDS) codes.

Undetected disk errors in RAID arrays

Though remarkably reliable, disk drives do fail occasionally. Most failures can be detected immediately; moreover, such, failures can be modeled and addressed using technologies such as RAID (Redundant Arrays of Independent Disks). Unfortunately, disk drives can experience errors that are undetected by the drive-- which we refer to as undetected disk errors (UDEs). These errors can cause silent data corruption that may go completely undetected (until a system or application malfunction) or may be detected by software in the storage I/O stack. Continual increases in disk densities or in storage array sizes and more significantly the introduction of desktop-class drives in enterprise storage systems are increasing the likelihood of UDEs in a given system. Therefore, the incorporation of UDE detection (and correction) into storage systems is necessary to prevent increasing numbers of data corruption and data loss events. In this paper, we discuss the causes of UDEs and their effects on data integrity. We describe some of the basic techniques that have been applied to address this problem at various software layers in the I/O stack and describe a family of solutions that can be integrated into the RAID subsystem.