db

1 files changed, 34 insertions, 6 deletions

diff --git a/db.json b/db.json
index e522471..5c1b943 100644
--- a/db.json
+++ b/db.json
@@ -2487,7 +2487,7 @@
     {
       "id": "find-your-face",
       "title": "Find Your Face",
-      "body": "Description of face recognition pipeline coming SOON!",
+      "body": "The Hansel & Gretel exhibition uses a custom facial recognition system that automatically captures visitors' facial biometrics as they move throughout the Armory. At all times, 5 surveillance cameras stream a 1080 pixel image at 10 frames per second (FPS) to a central image processing server located on site. Every frame from every camera is run through a series of computer vision algorithms that detect faces and convert them into biometric face prints. In less than 1 millisecond, using a minimum of 100 x 100 pixels, any face appearing in any surveillance feed is automatically extracted and stored in a local database as a 128-dimension feature vector.\r\n \r\nThese 128-dimension feature vectors represent what makes each face unique relative to other faces. Previous versions of facial recognition technology from the 1990s and early 2000s based recognition on seeing the face in a more human way, as linear combinations of composite photographs or by comparing the geometry between facial landmarks. Current state-of-the-art facial recognition algorithms, including the code used in the Hansel & Gretel system, make use of deep neural networks that automatically determine which combinations of facial features are the most useful for creating enough separation between between thousands of identities. In a controlled testing environment using frontal-pose images, the accuracy exceeds 99%. \r\n \r\nIn practice however, images captured in any surveillance system will typically be less accurate because of the differences between the pose, lighting, expression, optics and fashion accessories worn for the enrollment and capture photos. This effect is slightly minimized in the Hansel & Gretel facial recognition system because visitors are enrolled and matched on the same day. The accuracy of your matched photo will depend mostly on the equivalence of expression and pose between your submitted photo and your captured photos. According to a theory called the Biometric Zoo, match scores are also dependent on the relative similarity or difference of facial features amongst the enrolled population of faces. As with facial recognition technology and biometric recognition used in law enforcement and national security, no match is ever absolutely certain, but only a probabilistic determination based on the available matches in a database.\r\n \r\nIn the US the average person is caught on camera more than 75 times per day.  The fact that our identities are captured, analyzed, and stored is a fact of contemporary life that often goes unnoticed, but what are the implications this has on our society?",
       "disabled": false,
       "__index": 7,
       "dateCreated": "Sat, 20 May 2017 21:50:17 GMT"
@@ -2499,6 +2499,22 @@
       "disabled": false,
       "__index": 8,
       "dateCreated": "Mon, 22 May 2017 09:54:07 GMT"
+    },
+    {
+      "id": "curatorial-statement",
+      "title": "Curatorial Statement",
+      "body": "<i>Hansel & Gretel</i> is the latest work in the complex and exceptionally fruitful collaboration between Pritzker Prize winning Swiss architects Jacques Herzog and Pierre de Meuron (architects for the Armory’s renovation) and the Chinese artist/activist Ai Weiwei. Having worked together for fifteen years in the field of art and architecture, never defining their roles and thus creating unexpected results, they have collaborated on projects such as  the \"Bird's Nest\" stadium for the Beijing Olympic Games in 2008 and the 2012 pavilion for the Serpentine Gallery in London. <i>Hansel & Gretel</i> brings together their combined interests in the psychological impact of architecture and the politics of public space; creating a playful, strange and eventually eerie environment with different layers of reality revealed to the visitors first in the Drill Hall and then in the Head House of the Park Armory. \r\n\r\nThe initial impulse for the project was to transform the vast Thompson Drill Hall into a public park, a place of free movement and play, open 24 hours and accessible from street level. <i>Hansel & Gretel</i> is, however, quite the opposite, a dystopian forest of projected light where the floor rises up, as if lifted by an invisible force, and visitors are tracked by infrared cameras and surveyed by overhead drones as they systematically capture the park­goers' data and movements. Here the breadcrumbs of the famous Hansel and Gretel fairy tale are not eaten by birds but rather digital crumbs are gathered and stored, reminiscent of Ray Bradbury's poignant, 1953 science-fiction novel <i>Fahrenheit 451</i>, where an omniscient state surveils its citizens from the skies. \r\n\r\nEntering from a small street-level doorway on Lexington Avenue through a long darkening tunnel, the visitor experiences both psychological menace and exhilarating wonder upon exiting into the expansive landscape of the dimmed Drill Hall, animated by interactive projections mapping the visitor's every move. Utilizing state-of-the-art surveillance technology, the installation is both an enticingly playful and unnerving experience of what it means to be constantly watched, of public space without anonymity. Only upon leaving the Drill Hall and entering the hallways of the historic Head House, does the visitor discover through a continuation of the installation the extent of what has been seen and captured. An extensive digital library of surveillance histories and technologies is available for further research. \r\n\r\nIn an age of constant scrutiny and data storage beyond the knowledge and control of ordinary citizens, <i>Hansel & Gretel</i> is perhaps less fantastical and more menacing than it may at first appear. \r\n\r\nTom Eccles and Hans Ulrich Obrist, Curators \r\n",
+      "disabled": false,
+      "__index": 9,
+      "dateCreated": "Sat, 03 Jun 2017 19:09:34 GMT"
+    },
+    {
+      "id": "floor-projections",
+      "title": "Floor Projections",
+      "body": "Every movement in the Drill Hall is tracked and recorded from above by a grid of 56 small computers with infrared (IR) cameras, each attached to a projector suspended on a truss system high above the gently sloping floor . To allow for better tracking in low light conditions we installed an array of floodlights which operate in the near infrared spectrum and are thus is invisible to the human eye. Each computer runs a tracking algorithm that compares a preprogrammed image of the floor to the real time camera feed, and subtracts away any redundancy until only the difference features remain.  The moving person can thus be isolated, identified, and followed, which triggers a red rectangular outline and grid pattern to appear.  \r\n \r\nThe IR image now overlaid with tracking data undergoes a series of visual transformations. The powerful computers are able to make adjustments in cropping, scaling, warping, shifting, rotation, exposure, brightness, and sharpness, all within milliseconds.  Only then are the “real time” images sent to the projector.  The entirety of the 215 x 145 feet Drill Hall floor is covered by a total projection surface of 12360 x 7850 pixels, with each pixel being approximately ⅕ of an inch.  \r\n \r\nThe projections enable visitors to reflect upon the growing web of interconnected devices designed with ever evolving sensing capabilities and make tangible the reality that we leave digital traces almost everywhere.",
+      "disabled": false,
+      "__index": 10,
+      "dateCreated": "Sat, 03 Jun 2017 19:11:30 GMT"
     }
   ],
   "stream": [
@@ -2508,19 +2524,31 @@
       "streams": [
         {
           "text": "Entrance",
-          "uri": "https://www.youtube.com/watch?v=fO_-p8bi_Lk"
+          "uri": "https://www.youtube.com/watch?v=r72Azq0C0_E"
         },
         {
           "text": "Tunnel",
-          "uri": "https://www.youtube.com/watch?v=lDtcQID42Sg"
+          "uri": "https://www.youtube.com/watch?v=GON5zMPrG4M"
         },
         {
           "text": "Exit",
-          "uri": "https://www.youtube.com/watch?v=Szu3aBo8taQ"
+          "uri": "https://www.youtube.com/watch?v=LeKgQdZLJuQ"
+        },
+        {
+          "text": "Part 2 Left",
+          "uri": "https://www.youtube.com/watch?v=E_66XZcernM"
+        },
+        {
+          "text": "Part 2 Right",
+          "uri": "https://www.youtube.com/watch?v=Pm6kg0wJws0"
+        },
+        {
+          "text": "Infrared Cameras",
+          "uri": "https://www.youtube.com/watch?v=Kvo-_JxCj20"
         },
         {
-          "text": "Part 2 Entrance",
-          "uri": "https://www.youtube.com/watch?v=ErXmu-t6JcE"
+          "text": "Drones",
+          "uri": "https://www.youtube.com/watch?v=9GLDrLDALoo"
         }
       ],
       "disabled": false,