The Eyes

Robots permanently patrol the deep forests that surround Erewhon and separate the city from the rest of the world. There are not many of them. Invisibles wait for them, and whenever they hear the faint rumbling of the little machine, they emerge from the thicket where they'd taken refuge, plant themselves in the middle of the path, and wait with pounding heart: will the robot pass by without seeing them, extending their sorrow, or will the city finally recognize their face and outline, and welcome them back into its midst, endowing them once again with a first name, and a seat in the autom that will take them back home among their kin?


How beautiful are the roads that lead back to Erewhon after you’ve been forgotten by the city! The first suburbs appear, with people, stores, and cameras above...

Erewhon has eyes everywhere: in the streets, attached to lampposts, traffic lights, in front of banks and bus stops, at supermarket exits, inside homes…The images shot by domestic robots pass into the city’s brain, as do those recorded using mobile phones. Refrigerators, which also have an eye, smile at you, and call you by your name when you come close. There are caring cameras everywhere.

Erewhonians have just one fear, that the city’s eyes stop seeing them. They fear darkness, in which the eyes have difficulty detecting: they feel alone and abandoned. But that is nothing compared to the oblivion that the city can impose on them in broad daylight: by simply stopping to pay attention, no longer identifying their face or their gait, and by displaying welcoming emoticons on the screens they touch.

To attract the attention of Erewhon, inhabitants sometimes take the most dangerous tubes, roller coasters in which they slide at full speed, and spin out of control through a series of loops. They film themselves, telephone in hand, hoping to make the rounds of networks. Or much more perverse, they stop in front of one of these caring cameras. They simply want to be sure that their face appears on the control screen with their name. They may receive a fine for doing so, but no matter. This way they are sure they still have a place within the city’s brains.

Except that when such discourtesies accumulate, Erewhon forgets them. From one day to the next, or one minute to the next. It forgets them without warning or explanation. It forgets without reason. Maybe it’s even a bug, a malfunction. No one knows. All of a sudden, one is taboo: Invisible. Nothing more. Eyes no longer recognize you, doors remained closed before your face, machines resist your contact. And the most intimate of machines, your mobile phone, remains inert, no longer vibrating for you.

Adult computer vision experiment. [Source](
Adult computer vision experiment. [Source](

There are no other punishments in Erewhon, but there is no greater punishment for an Erewhonian than being abandoned by the city. Invisibles can’t even go home. So they take the path toward the large forests. By day and night, they walk in darkness, come to learn cold, hunger, fear of animals and, after dusk, true night, with no electrical light. Crackling noises in the thickets make them tremble. When Invisibles randomly run into each other along the paths, they don’t even think of speaking to each other. They listen for something else, the purring of the robot, which will identify them again as it wanders through the forest paths, marking the end of their punishment, and restoring them to life. Or in the end, they renounce this permanent expectation and penetrate deeper into the forest. They disappear. They've left Erewhon.

Films cités

Nikolai Smolyanskiy, Autonomous Drone Navigation with Deep Learning. Flight over 250 meter Forest Trail, 2017

Ngoc Anh Huynh, Lane Detection using Semantic Segmentation, 2017

Cory Lee, YOLOv2 Real-Time Object Detection on KITTI, 2016 - Native Video Deep Learning Demo, 2017

Bosch Security - Video analytics at the edge - Intelligent Tracking, 2016

Sergey Nuzhny, Multiple Objects Detection and Tracker, 2017

Denis Andreev, OpenCV multiple object tracking in realtime (children on playgound), 2016

Sakakibara Group, High-speed optical tracking of a flying insect, 2012

lfonso Pérez Escudero, Automatic tracking of 8 ants, 2013

Extended Demonstration Tracking Many Ants (Aphaenogaster Cockerelli), 2012

PeTrack Software, Pedestrian-dynamics experiment: pedestrian tracking (PeTrack) software, 2010

Chris Fotache, Land Features Detection on Drone Video with Deep Learning, 2017

FLIR Thermal Vehicle Detection, 2016

IPS Intelligent Video Analytics, IPS Loitering Detection Axis, 2017

Ipsotek, Smoke detection in tunnel, 2012

Seascom, Video IQ Intelligent Video Analytics Prevent Theft at Sheep Farm, 2017

Object detection using Yolo

Michael Virgo, Lane Detection with a Fully Convolutional Neural Network, 2017

Chapitre 3
Chapitre 5
The Blissful Ones