-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathcli.toml
149 lines (123 loc) · 6.2 KB
/
cli.toml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
# Enums
enums.feature_model = ["RESNET", "CM", "HOG", "AvgPool", "L3", "FC"]
enums.dimensionality_reduction_technique = ["SVD", "NNMF", "LDA", "kmeans"]
enums.label = ['gerenuk', 'hawksbill', 'headphone', 'ant', 'butterfly', 'lamp', 'strawberry', 'water_lilly', 'chandelier', 'dragonfly', 'crab', 'pagoda', 'dollar_bill', 'emu', 'inline_skate', 'platypus', 'dalmatian', 'cup', 'airplanes', 'joshua_tree', 'cougar_body', 'grand_piano', 'trilobite', 'brontosaurus', 'wild_cat', 'pigeon', 'dolphin', 'soccer_ball', 'wrench', 'scorpion', 'flamingo_head', 'nautilus', 'accordion', 'cougar_face', 'pyramid', 'camera', 'barrel', 'schooner', 'cellphone', 'panda', 'revolver', 'lobster', 'menorah', 'lotus', 'stapler', 'crocodile', 'chair', 'helicopter', 'minaret', 'starfish', 'ceiling_fan', 'ketch', 'mayfly', 'wheelchair', 'bass', 'yin_yang', 'crocodile_head', 'saxophone', 'beaver', 'mandolin', 'bonsai', 'Leopards', 'car_side', 'ibis', 'electric_guitar', 'kangaroo', 'stegosaurus', 'ferry', 'snoopy', 'umbrella', 'rhino', 'okapi', 'watch', 'brain', 'gramophone', 'scissors', 'rooster', 'cannon', 'binocular', 'anchor', 'octopus', 'buddha', 'laptop', 'windsor_chair', 'hedgehog', 'pizza', 'euphonium', 'stop_sign', 'Motorbikes', 'sea_horse', 'flamingo', 'ewer', 'garfield', 'crayfish', 'Faces_easy', 'Faces', 'sunflower', 'llama', 'elephant', 'tick', 'metronome']
enums.latent_semantics = ['ls1', 'ls2', 'ls3', 'ls4']
enums.classifier = ['NN', 'DT', 'PPR']
enums.relevance_feedback_systems = ['svm', 'prob']
# Removed 'BACKGROUND_Google' as all images have odd id ---> No image or label FDs
# Tasks
[tasks.0]
description = '''
Implement a program which computes and prints the “inherent dimensionality” associated with the even numbered Caltec101 images or each unique label of the even numbered Caltec101 images.
'''
usage = ""
[[tasks.0.args]]
id = "label"
type = "bool"
description = "Uses label vectors if this is specified"
[[tasks.0.args]]
id = "image"
type = "bool"
description = "Uses image vectors if this is specified"
[tasks.1]
description = '''
Implement a program which,
– for each unique label l, computes the corresponding k latent semantics (of your choice) associated with the even
numbered Caltec101 images, and
– for the odd numbered images, predicts the most likely labels using distances/similarities computed under the
label-specific latent semantics.
The system should also output per-label precision, recall, and F1-score values as well as output an overall accuracy
value.
'''
usage = ""
[[tasks.1.args]]
id = "k"
type = "int"
description = "top-k latent semantics"
[tasks.2]
description = '''
Implement a program which,
– for each unique label l, computes the correspending c most significant clusters associated with the even numbered Caltec101 images (using DBScan algorithm); the resulting clusters should be visualized both
∗ as differently colored point clouds in a 2-dimensional MDS space, and
∗ as groups of image thumbnails. and
– for the odd numbered images, predicts the most likely labels using the c label-specific clusters.
The system should also output per-label precision, recall, and F1-score values as well as output an overall accuracy
value.
'''
usage = ""
[[tasks.2.args]]
id = "c"
type = "int"
description = "C significant clusters"
[[tasks.2.args]]
id = "visualize"
type = "bool"
description = "To visualize the clusters"
[tasks.3]
description = '''
Implement a program which,
– given even-numbered Caltec101 images,
∗ creates an m-NN classifier (for a user specified m),
∗ creates a decision-tree classifier,
∗ creates a PPR based clasifier.
For this task, you can use feature space of your choice.
– for the odd numbered images, predicts the most likely labels using the user selected classifier.
The system should also output per-label precision, recall, and F1-score values as well as output an overall accuracy
value.
'''
usage = ""
[[tasks.3.args]]
id = "classifier"
type = "str"
enum = "classifier"
description = "type for classifier to be used"
[[tasks.3.args]]
id = "m"
type = "int"
description = "m nearest neighbors for NN classifier"
[[tasks.3.args]]
id = "p"
type = "float"
description = "random jump probability for PPR classifier"
[[tasks.3.args]]
id = "image_ids"
type = "str"
description = "Query image IDs separated by commas (no space)"
[tasks.4]
description = '''
- Implement a Locality Sensitive Hashing (LSH) tool (for Euclidean distance) which takes as input
(a) the number of layers, L
(b) the number of hashes per layer, h
(c) a set of vectors as input and creates an in-memory index structure containing the given set of vectors.
- Implement a similar image search algorithm using this index structure storing the even numbered Caltec101 images and a visual model of your choice (the combined visual model must have at least 256 dimensions): for a given query image and integer t
- Visualizes t most similar images
- outputs the number of unique and overall number of images considered during the process
For the relevance feedback system,
Let us consider the tag set “Very Relevant (R+)”, “Relevant (R)”, “Irrelevant (I)”, and “Very Irrelevant (I-)”. Implement
– an SVM based relevance feedback system,
– a probabilistic relevance feedback system
which enable the user to tag some of the results returned by 4b as and then return a new set of ranked results, relying on the feedback system selected by the user, either by revising the query or by re-ordering the existing results.
'''
usage = ""
[[tasks.4.args]]
id = "L"
type = "int"
description = "number of layers"
[[tasks.4.args]]
id = "h"
type = "int"
description = "number of hashes"
[[tasks.4.args]]
id = "t"
type = "int"
description = "t most similar images"
[[tasks.4.args]]
id = "image_id"
type = "int"
description = "Query image ID"
[[tasks.4.args]]
id = "rfs"
type = "str"
enum = "relevance_feedback_systems"
description = "Type of relevance feedback system"