Image sequence evaluation: 30 years and still going strong

Abstract

Over the past thirty years, technical means to record, digitize, store, process, and present video image sequences have multiplied in capacity b y several orders of magnitude. This large quantitative improvement is about t o turn into a qualitative change regarding the research to be pursued within this area in the future. Whereas up to now a n exploratory type of approach has prevailed, the increasingly available technical means for image sequence evaluation facilitate the investigation of large amounts of data. This requires to build encompassing systems and to study their behavior which in t u r n is expected to stimulate theoretical developments: even small effects of a n assumption can be studied in detail within a system environment, thereby enabling stepwise improvements of theoretical approaches underlying the construction of system solutions. The followi n g discussion will emphasize fundamental experimental and theoretical advances rather than details of particular approaches. 149 0-7695-0750-6/00 $10.00 0 2000 IEEE

Reasoning about Depth and Motion from an Observer's Viewpoint

“…On a broader sense, research on image sequence interpretation includes a myriad of different methods and techniques that have been proposed over the past thirty years. However, as surveyed in (Nagel, 2000), the problem of obtaining a robust automated description of observed scenes remains still open.…”

Markovian Energy-Based Computer Vision Algorithms on Graphics Hardware

“…This is the well known aperture problem [15]. In order to guaranty the uniqueness of a consistent solution, several approaches have been proposed [13]. Among these approaches, many add a regularization term (or smoothness constraints) whose essential role is to rightly constrain the ill-posed nature of this inverse problem.…”

“…The goal of motion estimation is to estimate the direction and magnitude of optical motion over each site s ∈ S of an animated sequence [13,14]. Among the solutions proposed in the literature, many are based on an hypothesis called lightness consistency.…”

Unified Target Detection and Tracking Using Motion Coherence

“…The problem of recovering the motion of potential targets from image sequences has been widely addressed [4,22]. General approaches to estimating target trajectories have been both token (e.g., [16,24,25]) and area (e.g., [2,9,18]) based.…”