关于of 后使用冠词的总结

关于of后使用冠词的总结：

一直想弄清楚，结果搜索了几次，都没有找到一个完整的解释。干脆自己总结了，有不妥之处，请指正！

Ⅰ. 一般在of后的名词不可数名词 要有“the”，泛指和专有名词除外（多用泛指）：

Segmentation is a by-productof recognition:

the paper of Russell et al

The statistics of the data,

consisting of the many variantsof spectral factorization

that can describe segments of complex shape and appearance这里of后没有冠词

the model of Fig. 1

the shape ofthe segment

the set ofDirichlet Process Gaussian mixture model (DP-GMM)

the level of confidence

the model of (5)

the SPMM of Section 2.1.3

the properties ofthe SPMM

an object ofcomplex shape (the stone arch),

the nonparametric approximationof x.

from averaging patches of different color but similar gradient pattern

the similarity of segmentation accuracy

a certain amount of supervision,

a certain amount of supervision is

the bottom three rows of Table 2

a drop of 0.16 in accuracy.

the use of human supervision

The lack of supervision

 

Ⅱ. 一般of后的单数可数名词，要有不定冠词a 或 定冠词 the， 除非是用了“each, every, that, my，both…”等，如：

 the shape and appearance of each segment

the components ofthe object

the appearance of an image

the segments of each image and the correspondence between such segments

Plate diagram ofourgenerative model

the number of segment？？？？

the nth observation of image m????（有问题？）

a building block of a larger model

The choice of a Dirichlet distribution

The advantage ofour model is

 

Ⅲ. 如果是复数家-s，前面可以不要定冠词“the”除非特别指定。如：

the popular “bag of visual words” model

a collection of images

a unified view of both tasks

a small dictionary of “textons,”

where statistics of words are used to

the visual appearance of different segments in images.

independently of the visual words

the shape of clusters

segments of complex shapes

the use of superpixels

the total number of pixels

all the images offaces,

the features that relate segments are restricted to size,shape, and average color ofthe segments.

 

 

 

Ⅳ. 特殊的一类固定搭配of后接单数，如：

this type of model is used

the variables of interest,

 

Ⅴ. 在of后接动名词，如

the idea of carrying out category learning for recognition and segmentation jointly.

The performance of fitting a GMM is of the lowest quality

The appearance of recurring elements

 

复合示例句子：

Our model represents each image as a composition of segments, where a segment could correspond to a whole object (e.g., a cow) or to a partof an object (e.g., a leg), to a patch of a distinctive texture, or to a “nonsense” homogeneous region in the background.

 

The more challenging problems of simultaneous segmentation and recognition of an entire image collection,…

The shape and appearance of a segment across a collection of imagesof objects…

以上是从一篇paper中摘录下来的，《Unsupervised Learning of Categorical Segments in Image Collections》PAMI2012

I couldn'tthink of the name of that man anyhow.  

我怎么也想不出那人的名字来了。

Such recurring segments may bethought of as the “parts” of corresponding objects that appear multiple times in the image collection.(考虑成…)