By Megan Chavez,2014-09-15 10:48
10 views 0

    听说6 076424 齐立荣

    Academic Paper Analysis

    This paper, named A Classified and Comparative Study of Edge Detection Algorithms, was published in Information Technology: Coding and Computing, 2002 IEEE. Now analyse the paper in the following aspects.

    ?. The brief outline of the paper

    1. Introduction

    First introduce the definition of edge detection. Then give a classification in brief. 2. Classification of Edge Detectors

    Introduce a new classification of edge detection algorithms in detail. ? Gradient edge detectors;

    ? Zero Crossing;

    ? Laplacian of Gaussian (LoG);

    ? Gaussian Edge Detectors;

    ? Colored Edge Detectors

    3. Analyzing Advantages and Disadvantages

    In theory, the advantages and disadvantages of some available algorithms within the introductive category are discussed.

    4. Comparison

    With implemented in C++, two sets of images resulting from the application of those algorithms are presented. Considering the criteria then analyze the performance of the algorithms in detail.

    5. Conclusions

    Subjective evaluation of images showed that under noisy conditions ISEF, Canny, Marr-Hildreth, Kirsch, Sobel, Lapla2 and Lapla1 exhibit better performances, respectively.

    6. References

    ?. Means of Paragragh Development

    1. Introduction

    1.1 The definition of edge detection


    听说6 076424 齐立荣

    “An edge is defined by a discontinuity in gray level values. In other words, an edge is

    the boundary between an object and the background.”———— by Definition.

    1.2 Classification in brief

    The Classification…… the following differentiation operators:

    ? Gradient edge Detectors (first derivative or classical) ? Zero crossing (second derivative)

    ? Laplacian of Gaussian (LoG)

    ? Gaussian edge detectors

    ? Colored edge detectors ———— by Division-and-Classification.

    2. Classification of Edge Detectors

    “Edge detectors may well be classified into 5 categories as follows:……”————

    by Division-and-Classification.

    3. Analyze Advantages and Disadvantages

    “In this respect, we first present some advantages and disadvantages of algorithms [1, 5, and 7] within the context of our classification in Table 1.”

    ————by Comparison-and-Contrast.

    4. Comparison

    4.1 Criteria:Precision,Resolusion and Accuracy

    In the past two decades several algorithms ……The availability of well-defined

    quality criteria is important ————by Cause-and-Effect.

    4.2 Relatated parameter:SNR and AVR

    “Table 2 describes the corresponding values of the SNR for each edge detector.”

    ————by Division-and-Classification.

    4.3 Observe and analyze the resultant images

    “Applying these seven operators to a noisy image shows that with noisy images, second derivative operators, like ISEF and Canny, exhibit better performance but require more computations because of smoothing an image with a Gaussian function first and then computing the gradient.” ————by Cause-and-Effect.

    4.4 Discuss the performance of algorithms

    ROC curve of the comparison is shown in Figure 3, Signal to Noise Ratio (SNR) of


    听说6 076424 齐立荣

    seven mentioned operators is illustrated.” ————by Comparison-and-Contrast.

    5. Conclusions

    In this paper we classified the most commonly used algorithms into five category, then seven

    algorithms have been applied to 30 images and lastly two sets were presented. Subjective evaluation of images ————by Time.

    ?. The characteristic of writing

    1. This paper is well organized,clear and easy to read.

    2. The aim of paper is clearly expressed in the introduction.

    3. One point leads logically to the next.

    4. Avoiding abuse of euphemism in the paper.

    5. Vary sentence length and structure.

    6. The article is more objective by using the third person.

    7. The use of more precise and formal vocabulary.

    8. Using more noun phrases to replace subordinate clauses.

    9. There is a clear conclusion.

    10. The avoidance of repeating the same words or the words similar in meaning. 11. Using a table for the purpose of comparison. As we all know visual presentationg of information aids the reader and contributes to the message of the document.A table is a convenient way to present a large amout of data in a small amout of space. ?. The important sentences

    Here are some important sentences?

    1. Since edge detection is in the forefront of image processing for object detection, it is crucial to have a good understanding of edge detection algorithms. ————以形


    2. It is shown that under noisy conditions, ISEF, Canny, Marr-Hildreth, Kirsch,Sobel, Lapla2, Lapla1 exhibit better performance,respectively.

    3. Seven edge detectors in the context of the above mentioned classification, which are more commonly in use, are selected and then tested. ————which 引导的非限定


    听说6 076424 齐立荣


    4. Since edge detection is in the forefront of image processing for object detection, it is crucial to have a good understanding of edge detection algorithms. ————原因


    5. The Classification that we introduce in this paper is based on the behavioral study of these edges with respect to the following differentiation operators?————主语


    6. In Marr-Hildreth, locality is not especially good and the edges are not always thin, still this edge detector is much better than the classical ones in cases of low signal to noise ratio followed by Kirsch, Sobel, Lapla2 and Lapla1 respectively. ————比较


    7. The importance of the classification is that it simplifies several problems in Artificial Vision and Image Processing, by associating specific processing rules to each type of edges [2]. ————表语从句

    8. As edge detection is a fundamental step in computer vision it is necessary to point out the true edges to get the best results from the matching process. ————原因状


    ?. The key words

    Following is all the key words.

    Edge Detection 边缘检测

    Algorithms 算法

    Image Processing 图像处理

    Signal to Noise Ratio?SNR?信噪比

    Zero Crossing 零交叉

    Classification 分类

    Gray level values 灰度值

    Boundary 边界

    Parameters 参数

    Gradient 梯度


    听说6 076424 齐立荣

    Detector 探测器

    Laplacian 拉普拉斯

    Gaussian 高斯

    Vector 向量

    Orientation 位置

    Pixel 像素

    Precision 精度

    Resolution 辨析率

    Accuracy 精确度

    ?.Idiomatic expressions

    1. 主谓搭配?The availability of well-defined quality criteria is important.

    2. 名词与名词?Edge Detection Algorithms?Gaussian filtering?machine vision

    3. 形容词与名词搭配?gray level values?classical operators?digital images

    4. 动宾搭配?extract edges?detect edges

    5. 短语动词?depends on?associat to?be derived from?be classified into?be divided

    into?point out?be applied to?with respect to?apply to

    ?.Methods of Citing Other’s Words

    The article cited the works of others in the definition and data.According to the order of bibliographies quoted in the paper, the paper indicates the number of the corresponding entries.For example, “The shape of edges in images depends on many

    parameters: The geometrical and optical properties of the object, the illumination conditions, and the noise level in the images [1].”

    ?. The Harvest

    This paper is mainly about the study of the edge detection algorithms.It is devided into four paragraphs,which we label in order introduction,classification,discussion and conclusion. The format of this paper is very standard.After reading this paper, I not only have got some professional knowledge, but also got many technologies of the writing skill of academic paper.

    By reading the paper, I have learned how to arrange paragraphs in the paper, and


    听说6 076424 齐立荣

    how to develop a paragraph with sentence. And it is also very important to know how to express in a formal way. In lectures,we will likely her less formal speech;however,in writing we should use a more formal form if one exists.A formal academic paper written in informal English may be considered too simplistic, even if the actual ideas and data are complex. And a academic paper must present facts in exectly the way as they are,free from any distortion.Any information conveyed must be easily understood with no risk of misdirection,ambiguity or wandering. To do so the paper must carry all the paragraphs as possible,no more,no less. Compared with general writing,passive voice is much more preferrable in academic writing,to appear objective and unemotional,to draw the readers’ attention to the process or result with no interest in who causes such a result or process. Academic writing is featured in grammer by frequent use of present tense to express something premanent, like characteristic, process, disposition, general truth or facts,because such information in academic paper usually requires no specific reference to time.Present tense is also used in definition,calssification or analysis of the experiments. But in describing an experiment or research process, past tense is preferable to show it was just tentative,maybe far from objective because of various factors unable to be predicted by the researchers.

    In conclusion, I have learned a lot about requirements of academic writing after analysing this paper, and I believe it will benefit me for a long time.


Report this document

For any questions or suggestions please email