首页
编程语言
数据库
网络开发
Algorithm算法
移动开发
系统相关
金融统计
人工智能
其他
首页
>
> 详细
代写program、C/C++程序语言代做
Coursework 2 (Group) – Scene Recognition
Brief
This is a group coursework: please work in teams of four people.
Due date: Wednesday 10th January, 16:00.
Development data download: training.zip in the coursework (CW) folder
Testing data download: testing.zip in the CW folder
Required files: report.pdf; code.zip; run1.txt; run2.txt; run3.txt
Credit: 25% of overall module mark
Overview
The goal of this project is to introduce you to image recognition. Specifically, we will examine the
task of scene recognition starting with very simple methods -- tiny images and nearest neighbour
classification -- and then move on to techniques that resemble the state-of-the-art.
This coursework will run following the methodology used in many current scientific benchmarking
competitions/evaluations. You will be provided with a set of labelled development images from
which you are allowed to develop and tune your classifiers. You will also be provided with a set of
unlabelled images for which you will be asked to produce predictions of the correct class.
Details
You will need to write software that classifies scenes into one of 15 categories. We want you to
implement three different classifiers as described below. You will then need to run each classifier
against all the test images and provide a prediction of the class for each image.
Data
The training data consists of 100 images for each of the 15 scene classes. These are arranged in
directories named according to the class name. The test data consists of 2985 images. All the
images are provided in JPEG format. All the images are grey-scale, so you don't need to consider
colour.
Objective measure
The key classification performance indicator for this task is average precision; this is literally the
proportion of number of correct classifications to the total number of predictions (i.e. 2985).
Run conditions
As mentioned above, you need to develop and run three different classifiers. We'll refer to the
application of a classifier to the test data as a "run".
Run #1: You should develop a simple k-nearest-neighbour classifier using the "tiny image" feature.
The "tiny image" feature is one of the simplest possible image representations. One simply crops
each image to a square about the centre, and then resizes it to a small, fixed resolution (we
recommend 16x16). The pixel values can be packed into a vector by concatenating each image
row. It tends to work slightly better if the tiny image is made to have zero mean and unit length.
You can choose the optimal k-value for the classifier.
Run #2: You should develop a set of linear classifiers (an ensemble of 15 one-vs-all classifiers)
using a bag-of-visual-words feature based on fixed size densely-sampled pixel patches. We
recommend that you start with 8x8 patches, sampled every 4 pixels in the x and y directions. A
sample of these should be clustered using K-Means to learn a vocabulary (try ~500 clusters to
start). You might want to consider mean-centring and normalising each patch before
clustering/quantisation. Note: we're not asking you to use SIFT features here - just take the pixels
from the patches and flatten them into a vector & then use vector quantisation to map each patch
to a visual word.
Run #3: You should try to develop the best classifiers you can! You can choose whatever feature,
encoding and classifier you like. Potential features: the GIST feature; Dense SIFT; Dense SIFT in a
Gaussian Pyramid; Dense SIFT with spatial pooling (commonly known as PHOW - Pyramid
Histogram of Words), etc. Potential classifiers: Naive bayes; non-linear SVM (perhaps using a linear
classifier with a Homogeneous Kernel Map), ...
Run prediction format
The predictions for each run must be written to a text file named runX.txt (where X is the run
number) with the following format:
For example:
...
0.jpg tallbuilding
1.jpg forest
2.jpg mountain
3.jpg store
4.jpg store
5.jpg bedroom
...
Restrictions
• You are not allowed to use the testing images for anything other than producing the final
predictions They must not be used for either training or learning feature encoding.
The report
The report must be no longer than 4 sides of A4 with the given Latex format for CW2, and must be
submitted electronically as a PDF. The report must include:
• The names and ECS user IDs of the team members
• A description of the implementation of the classifiers for the three runs, including information on
how they were trained and tuned, and the specific parameters used for configuring the feature
extractors and classifiers. We expect that your "run 3" section will be considerably longer than the
descriptions of runs 1 & 2.
• A short statement detailing the individual contributions of the team members to the coursework.
What to hand in
You need to submit to ECS Handin the following items:
• The group report (as a PDF document in the CVPR format same as CW2; max 4 A4 sides, no
appendix)
• Your code enclosed in a zip file (including everything required to build/run your software and to
train and use your classifiers; please don't include binaries or any of the images!)
• The run prediction files for your three runs (named "run1.txt", "run2.txt" and "run3.txt").
• A plain text file listing the user ids (e.g. xx1g20) of the members of your team; one per line.
Marking and feedback
Marks will be awarded for:
• Successful completion of the task.
• Well structured and commented code.
• Evidence of professionalism in implementation and reporting.
• Quality and contents of the report.
• The quality/soundness/complexity of approach used for run 3.
Marks will not be based on the actual performance of your approach (although you can expect to
lose marks if runs 1 and 2 are way off our expectations or you fail to follow the submission
instructions). We will open the performance rankings for run 3. !"#$
Standard ECS late submission penalties apply.
Individual feedback will be given to each team covering the above points. We will also give overall
feedback on the approaches taken in class when we announce the winner!
Useful links
• Matlab
o Image processing toolbox tutorials
o Recommended: VLFeat
§ Example of using VLFeat to perform classification
o Linear and non-linear SVMs
• Python
o numpy, PIL, sklearn (Scikit-learn), OpenCV, etc.
• C and C++
o OpenCV
o Recommended: VLFeat
o Example of using VLFeat to perform classification (Note this code is Matlab, but most of the
functionality is available in the C/C++ API)
• Java
o Recommended: OpenIMAJ
§ Chapter 12 of the tutorial deals with image classification
o BoofCV
Questions
If you have any problems/questions, use the Q&A channel on Teams
联系我们
QQ:99515681
邮箱:99515681@qq.com
工作时间:8:00-21:00
微信:codinghelp
热点文章
更多
辅导 cs1b spring 2024 tth hw...
2024-04-19
讲解 managing financial risk...
2024-04-19
辅导 cs 0449 – project 5: /...
2024-04-19
辅导 elec 2141 digital circu...
2024-04-19
讲解 csc171 — videogame pro...
2024-04-19
讲解 comp3411 artificial int...
2024-04-19
讲解 stat3061: random proces...
2024-04-19
辅导 accounting 452, spring ...
2024-04-19
辅导 finc5001 foundations in...
2024-04-19
辅导 7ssmm712 – topics in a...
2024-04-19
讲解 com 337 - film studies ...
2024-04-19
辅导 mes202tc - digital vlsi...
2024-04-19
辅导 geography 2041b distanc...
2024-04-19
辅导 ecos3006 international ...
2024-04-19
讲解 fit5225 2024 sm1 creati...
2024-04-19
讲解 cit 593: introduction t...
2024-04-19
讲解 math 4931: take home ex...
2024-04-19
辅导 csci 547|info 533: syst...
2024-04-19
辅导 cs536-s24 intro to pls ...
2024-04-19
讲解 fit5212 - assignment 1辅...
2024-04-19
热点标签
ecmt1010
engg1340
fit5216
cmpsc311
econ3208
basc0003
ee3121
acct2002
comp5313
busi2131
ise529
elec372/472
csit940/csit440
cenv6141
comp3027/comp3927
ftec5580
comp1433
msci223
mark203
en3098
eden1000
ece6483
econ4410
mats16302
com6521
can202
cs 6476
chc6186
comp3211
comp222
csc1002
comp10002
cs 161
comp282
swen20003
comp27112
comm1190
elec9764
acfi3308
acct7101
fin6035
comp2048
geog0163
comp2013
coen 146
dts101tc
comp4880/8880
cs 455
07
sehh2042
comp30023
stat0045.
fil-30023
celen085
psyc40005
math40082
are271
comp9311
ee5311
imse2113
comp 2322
acct2102
fnd109
int102
is3s664
is6153
data4000
accfin5034
fit5212
cs536-s24
fit5225
ecos3006
mes202tc
finc5001
stat3061
csc171
cs1b
7ssmm712
bu.450.760
comp3411
cs170
swen90004
cpt206
comp5313/comp4313—large
bl5611
kxo206
comp532
elec207
kxo151
cs 2820
cpt108
math2319
dts204tc
qm222
comp2511
ccs599
infs1001
mat2355
eeee4123
25721
ifn647
pols0010
hpm 573
cs 2550
comp9417
stat0023
csci 1100
qbus6860
comp2003j
cse340
cs 61b
cs360
fin 3080
ierg 4080
cs6238
cit 594
finm7406
hw6
elec9713
asb-2522
mso3610
lit301
mcd4540
geog0030
125.330
biol0006
125.320
cs3334
fit2093
acct1101
110.309
masy1-gc
cs314
elec0048
gds104
mg5637
fit2096
math5905
eel4837
sehs4515
cpt s 321
asb2522 investment
ma214
co2104
mgmt2015
32516
math32051
econ1012
mark2052
comp3310
econ0019
dsci 525
abmf3184
aps106
antc27
finm7401
itp122
tech2300
math3026
is2022
cao107
36318
cs 211
comp9024
fit1047
ics4u
2xc3
en.540.635
4qqmn506
finn3081
phys10362
sta601
ec481e
math5165
csi 2120
el1205
comp7250
ecos3013
beam065
info1113
comp2051
csc325
mne 6130
ai6126
inu1111
ecs150
is61x6
cse115
seng6110
bus265
cpts260
mphy0009
csc306
eco2011
ee3004
st332
idepg001
info6001
cpt106
finm7409
联系我们
- QQ: 99515681 微信:codinghelp
© 2024
www.7daixie.com
站长地图
程序辅导网!