Computer Vision - ACCV'98

These two volumes constitute the refereed proceedings of the Third Asian Conference on Computer Vision, ACCV'98, held in Hong Kong, China, in January 1998. The volumes present together a total of 58 revised full papers and 112 revised posters selected from over 300 submissions. The papers are organized in topical sections on biometry, physics-based vision, color vision, robot vision and navigation, OCR and applications, low-level processing, active vision, face and hand posture recognition, segmentation and grouping, computer vision and virtual reality, motion analysis, and object recognition and modeling.

On typical implementations of Hough transform for improving its performances.- Hierarchical segmentation and representation with dynamic link architecture neural network.- Perceptually consistent segmentation of texture using multiple channel filter.- Optimal edge detection under difficult imaging conditions.- Restoring image quality through structure preserving de-noising.- Feature saliency from noise variations in invariants.- Multiscale image representation and edge detection.- Rotation invariant texture features from Gabor filters.- Euclidean invariants of linear scale-spaces.- Segmenting objects at multiple scales: A robust approach.- Multi-grid edge models for magnifying digital images.- Scale and rotation invariant recognition method using higher-order local autocorrelation features of log-polar image.- Script and language identification from document images.- Document categorization for document image understanding.- Recognition of various bar-graph structures based on layout model.- Word-class bigram statistics language model for a hand-written chinese character recognizer.- Log classification by single X-ray scans using texture features from growth rings.- Precise and fast form identification method by using adaptive base lines for matching.- Combinatorial coarse classification method for OLCCR.- Detecting characters in grey-scale scene images.- Conic based image transfer for 2-D Objects: A linear algorithm.- Minimal conditions on intrinsic parameters for Euclidean reconstruction.- Surface based hypothesis verification in intensity images using geometric and appearance data.- Next best viewpoint (NBV) planning for active object modeling based on a learning-by-showing approach.- Object recognition by matching symbolic edge graphs.- Interpretation of complex scenes using Bayesian networks.- Recognition of urban scene using silhouette of buildings and city map database.- A cooperative inference mechanism for extracting road information automatically.- Model-based active object recognition using MRF matching and sensor planning.- Improved image classification using morphing.- Reconstruction of non-manifold objects from two orthographic views.- 3D object recognition using segment-based stereo vision.- The state of color vision research.- Color vision and color media processing research in Asia.- Recent advances in detection and description of buildings from multiple aerial images.- Visual surveillance of human activity.- Bayesian paradigm for recognition of objects — Innovative applications.- Toward motion picture grammars.- Hierarchical texture segmentation.- Range image segmentation: Adaptive grouping of edges into regions.- Optimising the complete image feature extraction chain.- A unified framework for salient curves, regions, and junctions inference.- Learning multiscale image models of 2D object classes.- 3D model centered framework for CV and VR.- Image-based geometrically-correct photorealistic scene/object modeling (IBPhM): A review.- Measuring object surface shape and reflectance properties.- Robust image composition algorithms for augmented reality.- Context-based recognition of manipulative hand gestures for human computer interaction.- An algorithm for recursive structure and motion recovery under affine projection.- Relative affine depth: Structure from motion by an uncalibrated camera.- The eigenspace method for rigid motion recovery from less than eight point correspondences.- 3D shape and motion analysis from image blur and smear: A unified approach.- 3D line's extraction from 2D Spatio-Temporal Image created by sine slit.- Toward non-intrusive motion capture.- Appearance based visual learning and object recognition with illumination invariance.- Evidence-based scene interpretation considering subjective certainty of recognition.- Robust hypothesis verification for model based object recognition using Gaussian error model.- Shape modeling from multiple view images using GAs.- 3-D reconstruction of multipart self-occluding objects.- On analysis of cloth drape range data.- VR models from epipolar images: An approach to minimize errors in synthesized images.- Shape and pose parameter estimation of 3D multi-part objects.- Generating 3D models of objects using multiple visual cues in image sequences.- Strategical tracking of polyhedral objects by reactive change of projection pattern - Reactive Range Finder -.- Autonomous vision-guided robot manipulation control.- A new adaptive approach on rapid obstacle detection in range image.- Recognition of shape models for general roads.- Visual detection of obstacles assuming a locally planar ground.- Potential-based modeling of 2D regions using non-uniform source distributions.- A linear algorithm for motion from three weak perspective images using Euler angles.- On Learning spatio-temporal relational structures in two different domains.- An efficient iterative pose estimation algorithm.- A new multistage approach to motion and structure estimation by gradually enforcing geometric constraints.- Tracking a person with pre-recorded image database and a pan, tilt, and zoom camera.- Recovery of motion and structure from optical flow under perspective projection by solving linear simultaneous equations.- Vector coherence mapping: A parallelizable approach to image flow computation.- Robust motion segmentation using rank ordering estimators.- Optical flow in the scale space.- Motion detection in temporal clutter.- A novel fast three-step search algorithm for block-matching motion estimation.- Moving vehicle detection and tracking in image sequences.- Gesture recognition from image motion based on subspace method and HMM.- Identifying faces under varying pose using a single example view.- Multiple camera based human motion estimation.- An autonomous facial caricaturing based on a model of visual illusion — Experimental modeling of visual illusion —.- 3D estimation of facial muscle parameter from the 2D marker movement using neural network.- Appearance-based face recognition under large head rotations in depth.- Skin-color modeling and adaptation.- Human information retrieval by face extraction and recognition on TV news images using subspace method.- Converting facial expressions using recognition-based analysis of image Sequences.- Muscle-based feature models for analyzing facial expressions.- A morphological method for moving object segmentation and posture recognition.- Detection of glasses in facial images.- Non-monotonic continuous dynamic programming for spotting recognition of hesitated gestures from time-varying images.- Face recognition using a face-only database: A new approach.