We address the problem of recognizing free-form 3D objects from a single 2D intensity image. A model-based solution within the alignment paradigm is presented which involves three major schemes { modeling, matching, and indexing. The modeling scheme constructs a set of model aspects which can predict the object contour as seen from any viewpoint. The matching scheme aligns the edgemap of a cand...

METHOD MAE MSD RODRIGUEZ ET AL. 655.7 697.8 LEMPITSKY ET AL. 493.4 487.1 ZHANG ET AL. 467.0 498.5 IDREES ET AL. 419.5 541.6 ZHANG ET AL. 377.6 509.1 CCNN 488.67 646.68 HYDRA 2S 333.73 425.26 HYDRA 3S 465.73 371.84 The Hydra model uses a pyramid of input patches cropped from the center of the target patch to provide multiscale information to the network. The counting by regression model with dee...

This paper presents a method for statically verifying that functions do not produce side eeects, in an object-oriented language. The described model, although not allowing any changes to pre-existing objects during a function call, permits an imperative style of programming, where new objects can be freely created and manipulated.

This paper presents an object recognition module development. This module uses a local feature approach to identify keypoints in free form objects and an unsupervised artificial neural network (ANN) to associate the nearest ones and get clusters of each object learned. The module uses A-KAZE feature descriptor and Growing Cell Structure (GCS) ANN. The module is validated using an own data base,...

Recognition of free-form objects in range data is hampered by the diiculty of extracting robust and reliable features. This paper develops two approaches for detection of local features for recognition from both polyhedral mesh approximations to free-form geometry and from range data. Several experiments with models of varying complexity are proposed and speciic recommendations are given for ap...

Instance-level object segmentation is an important yet under-explored task. The few existing studies are almost all based on region proposal methods to extract candidate segments and then utilize object classification to produce final results. Nonetheless, generating accurate region proposals itself is quite challenging. In this work, we propose a Proposal-Free Network (PFN ) to address the ins...

Most previous methods for tracking of multiple objects follow the conventional “tracking by detection” scheme and focus on improving the performance of category-specific object detectors as well as the between-frame tracklet association. These methods are therefore heavily sensitive to the performance of the object detectors, leading to limited application scenarios. In this work, we overcome t...

This thesis addresses problems of free-form object matching for the point vs. NURBS surface and the NURBS surface vs. NURBS surface cases, and its application to copyright protection. Two new methods are developed to solve a global and partial matching problem with no a priori information on correspondence or initial transformation and no scaling effects, namely the KH and the umbilic method. T...

This paper describes a new parallel parsing scheme for context-free grammars and our experience of implementing this scheme, and it also reports the result of our simulation for running the parsing program on a massive parallel processor. In our basic parsing scheme, a set of context freegrammar :,:ules is represented by a network of processorlike computing agents each having its local memory. ...

In this contribution we introduce a new model-free method for object tracking. The tracking is posed as a segmentation problem which we solve using the watershed algorithm. A framework is defined to compute the required topographic surface from distances to the predicted contour, intensity edges and motion edges. This multifeature tracking approach yields accurate results in the presence of obj...