This document discusses visual search and recognition, specifically large scale instance search. It outlines detecting objects across different images despite changes in scale, viewpoint, lighting and occlusion. Key steps include covariant feature detection to find corresponding regions, and generating invariant descriptors like SIFT to match features between images. The goal is to cast object recognition as nearest neighbor matching or text retrieval to perform efficient search over very large datasets.
28. 亠于亠 舒弍仂
J. O'Rourke and N. Badler. Model-based image analysis of
human motion using constraint propagation. IEEE Trans.
on Pattern Analysis and Machine Intelligence. 1980.
120. 3D 亳亰 2D
Y 仗舒舒仄亠 3D 仗仂亰, X 仗舒舒仄亠 2D
仗仂亰. 亢仆仂 仂亠仆亳 P(Y|X), 亳 亰舒亠仄 亞仍舒亟亳 仗仂
于亠仄亠仆亳 于 于亳亟亠仂
仍 仂亞仂 仗仂 于弍仂从亠 (X,Y) 仆舒仗礆 仂亠仆亳仄
P(Y|X) 仗仂仄仂 束Mixture of experts損
L. Sigal and M. J. Black. Predicting 3D people from 2D pictures. In AMDO 2006.
128. 个仂仄舒 亳 仗仂亰舒 亳亰 仂亟仆仂亞 亳亰仂弍舒亢亠仆亳
Peng Guan Alexander Weiss Alexandru O. Balan Michael J. Black,
Estimating Human Shape and Pose from a Single Image, ICCV 2009
129. 亳弍 亳亰 3D 仍亠亢亠仆亳
L. Sigal, D. Fleet, N. Troje, M. Livne Human Attributes from 3D Pose
Tracking,, European Conference on Computer Vision, ECCV 2010.