determine landmarking data model according to standard numbering system

[boazsender]

[tkellen]

Here is a proposed data model in "shorthand" notation:

-- all people who annotate images
annotater (id, city)

-- all possible images to be annotated
image (id, url, width, height)

-- all possible annotation types for an image (e.g. age)
annotation_type (id, name)

-- all options for a given annotation type
annotation_option (id, perception_type_id, name)

-- actual answers from annotaters classifying images
image_annotation (image_id, annotater_id, annotation_option_id)

-- all possible landmarks and the region they belong to (if they do)
landmark (id, numbering, region)

-- actual x/y coordinates from annotaters about each possible landmark for a given image
image_landmark_annotation (image_id, annotater_id, landmark_id, x, y)

-- "ground truth"
-- 4 or more people agreeing within a specified tolerance about the x/y
-- coordinates of a given landmark

The same data model, in full (untested) SQL:

CREATE TABLE annotater (
  id SERIAL PRIMARY KEY,
  city TEXT
);

CREATE TABLE image (
  id SERIAL PRIMARY KEY,
  url TEXT,
  width INTEGER,
  height INTEGER
);

CREATE TABLE annotation_type (
  id SERIAL PRIMARY KEY,
  name TEXT
);
INSERT INTO annotation_type (name) VALUES
  ('Perceived Age'),
  ('Perceived Gender'),
  ('Perceived Skintone');

CREATE TABLE annotation_option (
  id SERIAL PRIMARY KEY,
  annotation_type_id INTEGER NOT NULL REFERENCES annotation_type(id),
  name TEXT NOT NULL
);
INSERT INTO annotation_option (annotation_type_id, name) VALUES
  (1, 'infant'),
  (1, 'child'),
  (1, 'young adult'),
  (1, 'adult'),
  (1, 'elderly'),
  (2, 'other'),
  (2, 'female'),
  (2, 'male'),
  (3, 'skintone1'),
  (3, 'skintone2'),
  (3, 'skintone3'),
  (3, 'skintone4'),
  (3, 'skintone5');

CREATE TABLE image_annotation (
  id SERIAL PRIMARY KEY,
  annotater_id INTEGER NOT NULL REFERENCES annotater(id),
  image_id INTEGER NOT NULL REFERENCES image(id),
  annotation_option_id INTEGER NOT NULL REFERENCES annotation_option(id)
);

CREATE TABLE landmark (
  id SERIAL PRIMARY KEY,
  numbering INTEGER,
  region INTEGER
);
INSERT INTO landmark (numbering, region) VALUES
  (1, null),
  (2, null),
  (3, null),
  (4, null),
  (5, null),
  (6, null),
  (7, null),
  (8, null),
  (9, null),
  (10, null),
  (11, null),
  (12, null),
  (13, null),
  (14, null),
  (15, null),
  (16, null),
  (17, null),
  (18, null),
  (19, null),
  (20, null),
  (21, null),
  (22, null),
  (23, null),
  (24, null),
  (25, null),
  (26, null),
  (27, null),
  (28, 1),
  (29, 1),
  (30, 1),
  (31, 1),
  (32, 1),
  (33, 1),
  (34, 1),
  (35, 1),
  (36, 2),
  (37, 2),
  (38, 2),
  (39, 2),
  (40, 2),
  (41, 2),
  (42, 2),
  (43, 3),
  (44, 3),
  (45, 3),
  (46, 3),
  (47, 3),
  (48, 3),
  (49, 4),
  (50, 4),
  (51, 4),
  (52, 4),
  (53, 4),
  (54, 4),
  (55, 4),
  (56, 4),
  (57, 4),
  (58, 4),
  (59, 4),
  (60, 4),
  (61, 4),
  (62, 4),
  (63, 4),
  (64, 4),
  (65, 4),
  (66, 4),
  (67, 4),
  (68, 4);

CREATE TABLE image_landmark_annotation (
  id SERIAL PRIMARY KEY,
  image_id INTEGER NOT NULL REFERENCES image(id),
  landmark_id INTEGER NOT NULL REFERENCES landmark(id),
  annotater_id INTEGER NOT NULL REFERENCES annotater(id),
  x INTEGER NOT NULL,
  y INTEGER NOT NULL
);

[gnarf]

Thanks @tkellen - I think this gets us almost all the way there already! :)