summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authoradamhrv <adam@ahprojects.com>2019-06-27 23:56:43 +0200
committeradamhrv <adam@ahprojects.com>2019-06-27 23:56:43 +0200
commit5d1563ce4a3901d8fc86cb3a9867d4503bff9f05 (patch)
tree11fda72ace4460b56066635355b5d06b40a69490
parentc8be69355d562e8e3ff24bd76ca42ae4a492cdab (diff)
udpate todo
-rw-r--r--docs/interviews.md12
-rw-r--r--docs/notes.md1
-rw-r--r--docs/todo.md54
3 files changed, 36 insertions, 31 deletions
diff --git a/docs/interviews.md b/docs/interviews.md
new file mode 100644
index 00000000..0227ba0d
--- /dev/null
+++ b/docs/interviews.md
@@ -0,0 +1,12 @@
+# Interviews
+
+## OneZero
+
+> why havne't tech companies with unlimited resources paid for people's images to create a balanced and equitable dataset?
+
+
+> What makes the MS Celeb dataset different from a dataset of people's houses taken without their consent?
+
+
++
+Taking pictures isn't the problem. But pictures \ No newline at end of file
diff --git a/docs/notes.md b/docs/notes.md
index df71a14d..9055c78d 100644
--- a/docs/notes.md
+++ b/docs/notes.md
@@ -1,3 +1,4 @@
# Notes
+- LFW results page mentions two companies that used MS Celeb http://vis-www.cs.umass.edu/lfw/results.html can we add this to the MS Celeb verified list?
- https://verify.megapixels.cc/paper/ucf_crowd/verify/bfdc7cde3bbfcba738a5eefe9143417ebf7d8f5c mentions the UCF-QNRF dataset that was built searching Flickr/Google/etc using the words 'HAJJ,SPECTATOR CROWD,PILGRIMAGE,PROTEST CROWDandCONCERT CROWD.' \ No newline at end of file
diff --git a/docs/todo.md b/docs/todo.md
index 57dce0ca..12a3c495 100644
--- a/docs/todo.md
+++ b/docs/todo.md
@@ -7,16 +7,8 @@
## Global
-- JL: mobile CSS
- - lightbox/modal on mobile, close button not visible
-- AH: add social share graphic for homepage
- AH: add press kit/downloads
-## Splash
-
-- AH: create high quality 3d heads
-- JL/AH: add IJB-C names to word cloud
-
## Datasets
- JL: this paper isn't appearing in the UCCS list of verified papers but should be included https://arxiv.org/pdf/1708.02337.pdf
@@ -33,25 +25,17 @@
Collect Flickr IDs and metadata for:
- adience
-- flickr_faces
+- ffhq
- geofaces
-- h3d
- helen
- ibm_dif
-- images_of_groups
-- leeds_sports_pose
-- leeds_sports_pose_extended
- lfpw
- mafa
- me_glass
- megaage
- megaface
-- moments_in_time
- pipa
-- stickmen_family
- tiny_faces
-- ucf_crowd
-- used
- voc
- who_goes_there
- yfcc_100m
@@ -59,24 +43,32 @@ Collect Flickr IDs and metadata for:
## MSC Analysis:
-- [x] Adience, 97
-- [ ] LFPW, 9 completed, estimated 180
-- [ ] 300W, 3, estimated 150
+Flickr
+
+- [x] Adience, 118
+- [ ] LFPW, 146
+- [ ] 300W, 5
- [x] GeoFaces, 2
-- [ ] h3d, low hit-rate, maybe skip
-- [ ] Helen, 206, waiting for s2q cleanup
-- [x] MegaFace, 40
-- [ ] PIPA, 29 done, estimated 330
-- [ ] UCF-CC-50, 3, estimated 30, maybe skip
+- [ ] Helen, 283
+- [x] MegaFace, 53
+- [ ] PIPA, 35
+- [ ]
-subtotal: about 350
+## Add
+-
+ VGG Face
+- VGG Face 2
+- IMDB Face
+- IMDB Wiki
+- CelebA
+- CACD
+- FaceScrub
+- PubFig
+- PubFig83
+- UTK Face
+- LFW
-## FT Analysis:
-- [x] Brainwash
-- [x] Duke MTMC
-- [x] UCCS
-- [x] MSCeleb
- [ ] IJB-C (and IJB-A/B?)
- [ ] HRT Transgender
- [x] Town Centre